Mental health trajectories of women experiencing differing patterns of intimate partner violence across the first 10 years of motherhood

Intimate partner violence is commonly experienced by women and associated with poor mental health outcomes. Evidence regarding the patterns of IPV across time, and the associated long term course of depressive symptoms is lacking. The current study aimed to (a) identify patterns of physical and emotional IPV experienced by women over the 10 years since having their first child, and (b) identify trajectories depressive symptoms across the 10-year period for each pattern of IPV exposure. Data was drawn from the Mothers ’ and Young People ’ s Study (MYPS), a longitudinal study of 1507 mothers and their first born child. Data was collected during pregnancy and at one, four, and ten years postpartum. Using Latent Class Analysis, four distinct classes of IPV were identified: (1) Minimal IPV, (2) Early IPV exposure, (3) Increasing IPV, and (4) Persistent IPV. Latent growth modelling revealed that all classes characterised by some IPV exposure experienced elevated trajectories of depressive symptoms in comparison to the minimal IPV class. Those experiencing increasing and persistent IPV experienced the most severe course of depressive


Introduction
One in three women experience violence from an intimate partner during their lifetime (Brown et al., 2020;WHO, 2013). Intimate partner violence (IPV) is defined as any behaviour from a current or previous intimate partner that causes physical, psychological, or sexual harm. It can include acts of physical aggression (e.g., punching, slapping, pushing), psychological aggression (e.g., threats, humiliation), sexual violence (e.g., unwanted sexual contact, forced or coerced sexual acts), and patterns of controlling and/or coercive behaviour (e.g., restricting access to finances or social networks) (WHO, 2013). IPV disproportionally affects women and children and is considered a global health priority (WHO, 2021). It is well established that women who have experienced IPV are at substantially increased risk of physical and mental health difficulties compared to women who have not experienced IPV (Bacchus et al., 2018;Brown et al., 2020;Devries et al., 2013;Lövestad et al., 2017;Trevillion et al., 2012). Although there is growing interest in the long-term mental health outcomes associated with women's experiences of IPV (Bacchus et al., 2018;Herbert et al., 2022), few longitudinal studies have been conducted to examine the long-term patterns of exposure to IPV and the course of mental health symptoms over time. Investigating the long-term mental health outcomes associated with IPV is essential to guide policy and practice responses to support women and children.
Although women are likely to experience IPV across a number of years (FitzPatrick et al., 2022b;Thompson et al., 2006), there is limited evidence regarding the occurrence of different types of IPV over time. An Australian longitudinal study of over 1500 first time mothers conducted by our team assessed the occurrence of physical and/or emotional abuse across a ten-year period post childbirth (FitzPatrick et al., 2022b). This study found that while the overall prevalence of IPV remained consistent at each timepoint, there was substantial variation in women's individual experiences of physical and emotional IPV across the ten years. Women who experienced combined physical and emotional IPV were most likely to have reported IPV at multiple timepoints. Emotional IPV alone was the most prevalent type of IPV, reported by a quarter of study participants across the study period. In contrast, just 5% reported physical IPV alone (FitzPatrick et al., 2022b). A small number of other studies have investigated the trajectories of IPV experiences over time, frequently describing these as 'increasing ', decreasing' or remaining 'stable' (Poehacker et al., 2020;Swartout et al., 2012). For example, one study of 1408 women in South Africa found no change in psychological or physical IPV over time (Groves et al., 2014), however this follow-up was limited to the pregnancy and postnatal period. A more recent study conducted in the United States found five distinct patterns of psychological IPV experienced by mothers of school aged children (N = 1091) across an eight-year period (Poehacker et al., 2020). Specifically, the study found that although majority of women were characterised by experiencing a low stable risk of IPV, others experienced (a) low increasing, (b) moderate decreasing, (c) high decreasing, or (d) consistently elevated IPV. Despite these advances in understanding patterns of IPV, these studies have largely focused on either physical (Swartout et al., 2012) or emotional (Poehacker et al., 2020) IPV in isolation, rather than investigating how these co-occur across time.
The detrimental impact of all forms of IPV on women's mental health is well established (Bacchus et al., 2018;Devries et al., 2013;Lagdon et al., 2014). Research suggests that women who experience frequent forms of combined physical, sexual, and/or emotional IPV report the worst mental health outcomes (Brown et al., 2020;Hegarty et al., 2013;Lövestad et al., 2017). However, women experiencing emotional abuse alone are also at risk (FitzPatrick et al., 2022a;Lövestad et al., 2017;Woolhouse et al., 2012). A recent systematic review identified eight longitudinal studies investigating the association between IPV on subsequent depressive symptoms, with all eight studies identifying a positive association (Bacchus et al., 2018). Other research has sought to identify whether mental health may improve after leaving or seeking help for an abusive relationship. However, findings have been mixed. For example, a study of 406 women recruited from domestic violence services in the United States identified a significant decrease in depressive symptoms one year after accessing the service (Dutton et al., 2004). Similarly, a smaller study conducted in the United States (N = 141) (Campbell and Soeken, 1999) found an initial improvement in women's depressive symptoms following cessation of IPV. However, when assessed again at a 3-year follow-up, this study found that symptoms had returned to baseline levels (Campbell and Soeken, 1999). Other studies have also found that mental health symptoms may persist in the absence of continued exposure to IPV (Ford-Gilboe et al., 2009;Chuang et al., 2012). A key limitation of much of this research is its reliance on assessing IPV at a single timepoint and subsequent mental health (Ahmadabadi et al., 2019;Bacchus et al., 2018;Coker et al., 2002;Flach et al., 2011;Herbert et al., 2022;Wathen et al., 2016).
Few longitudinal studies have investigated the impact of IPV occurring over time on women's health. In a population-based study of Australian mothers (N = 1507), IPV at a single time point during the first four years following birth of a first child was associated with a 1.5-2.5 fold increase in likelihood of depressive symptoms at 10 years postpartum (Brown et al., 2021). The likelihood of depressive symptoms was higher (OR=3.4) for women who reported IPV at two time points within the first 4 years postpartum (Brown et al., 2020) and highest (OR= 2.9-4.5) for those women reporting recent IPV at the 10-year follow up. This study has also sought to investigate the longer-term relationship between different forms of IPV and women's mental health (FitzPatrick et al., 2023). Although current IPV was most strongly associated with depressive symptoms, this study found that women who experienced emotional IPV alone or combined physical and emotional IPV in the first and/or fourth year postpartum were at a 2-fold risk of experiencing depressive symptoms at 10 years postpartum (FitzPatrick et al., 2023). This research demonstrates the considerable impact of IPV on women's mental health, however, does not determine how women's mental health might change over time within the context of different patterns of exposure to IPV.
There is currently limited research investigating the mental health outcomes associated with patterns of IPV over time. One recent Australian study of women (N = 548) attending a general medical clinic with depressive symptoms, identified five distinct trajectories of IPV across a four-year period . This study found poor mental health outcomes four years after enrolment for women who reported moderate persistent, as well as high and increasing levels of IPV. Conversely, women who reported relatively high levels of IPV at baseline but little or no IPV by four years, had similar mental health outcomes to those experiencing minimal or no IPV across all timepoints . These findings provide important insights into the mental health trajectories of women who experience IPV within a clinical sample. Given that many women who experience IPV and poor mental health do not seek support (Gartland et al., 2022), there is a need to understand patterns of both IPV and mental health within broader community-based samples. Moreover, given the additional challenges for women experiencing IPV who are also raising children (Fogarty et al., 2019;Vatnar and Bjørkly, 2010), it is also important to investigate IPV and the long-term course of mental health symptoms for women who are also mothers. This is particularly important given that mental health difficulties during the postnatal and early parenting period are common (Gavin et al., 2005;Howard et al., 2014;Woolhouse et al., 2015), and that this period often coincides with an age where risk of IPV for women is also at its highest (i.e., 18-34 years; AIHW, 2022).

Study aims
In summary, there is still much to understand about how patterns of emotional and physical IPV are experienced by women across time, and how this relates to mental health over time. Understanding how mental health changes over time in the context of patterns of exposure to different types of IPV can assist in guiding targeted mental health support to prevent the onset of severe symptoms for women and support their healing and recovery. The current study used data drawn from the Mothers' and Young People's Study (MYPS) to investigate patterns of IPV and the course of depressive symptoms amongst a community sample of women across the 10 years after giving birth to their first child. MYPS collected data on women's experiences of IPV and depressive symptoms from pregnancy to 10 years postpartum, providing an opportunity to assess patterns of emotional and physical IPV experienced by women over time and the associated mental health trajectories. The specific aims of this paper were to (1) identify patterns of physical and emotional IPV experienced by women across the 10 years after giving birth to their first child, and (2) identify trajectories of depressive symptoms across the 10-year period for each pattern of IPV exposure.

Study design
Data were drawn from the Mothers' and Young People's Study (previously the Maternal Health Study), a longitudinal study of the health and wellbeing of 1507 new mothers and their first born child. Women who were registered to give birth across six metropolitan public hospitals within Melbourne, Victoria were invited to take part in the study. Recruitment took place between 2003 and 2005 and women were eligible to take part if they were (a) nulliparous, (b) 18 years or older, (c) ≤ 24 weeks gestation, and (d) proficient in English language. Hospital staff mailed study packs (study information, consent form, baseline questionnaire, reply paid envelope) to eligible women inviting them to participate. Women who had not responded after two weeks were mailed a reminder card. Follow up data were collected at 3 months, 1, 4 and 10 years postpartum. Further details on study design are outlined in the study protocol (Brown et al., 2006) and updated study profile (Brown et al., 2021).
A total of 1507 women were deemed eligible to participate and enroled in the study. Although a precise response rate is unable to be determined (due to many women likely receiving invitations via more than one pathway), we conservatively estimate that 33% of eligible women enroled in the study. Participant retention was 90% at 1 year, 73.1% at 4 years and 63.2% at 10 years postpartum. Selective attrition was observed with participants who completed the 10-year follow-up being more likely to be older, born in Australia, have obtained post high school qualifications, and less likely to report depressive symptoms and IPV in the first year postpartum.

Demographic and social health information
Demographic and social health information collected at baseline in early pregnancy and at each of the follow-up time points in early pregnancy included maternal age, country of birth, language spoken at home, employment status, holding a government health care concession card (a marker of socioeconomic disadvantage), educational level, and relationship status.

Intimate partner violence
Exposure to emotional and physical IPV was assessed using the Composite Abuse Scale (CAS; Hegarty et al., 2005) at 1, 4 and 10 years postpartum. Women were asked how frequently they had experienced physical IPV (e.g., Slapped me, beat me), or emotional IPV (e.g., Told me I was ugly, told me I wasn't good enough) from a current or former partner during the previous 12-month period. The short 18-item version was used at 1, 4, and 10 years postpartum. Items were rated on a 5-point Likert scale (never, only once, several times, once per week, or daily) with 11 items asking women about their experiences of emotional IPV, and seven related to physical IPV. Women were asked to respond whether they were in a current relationship or not. In line with the scale guidelines, women were categorised as experiencing emotional IPV if they endorsed a score of ≥3 on emotional IPV items, or ≥1 on physical IPV items. The CAS has demonstrated excellent validity in identifying physical and emotional IPV (Hegarty et al., 2005).

Depressive symptoms
Maternal depressive symptoms were assessed at 3 months, 1 year, 4 years and 10 years postpartum using the 10-item Edinburgh Postnatal Depression Scale (EPDS; Cox et al., 1987). The EPDS asks women to indicate the extent to which they have been experiencing depressive symptoms across the last week on a 4-point scale ranging from 0=never to 3= most of the time. Examples of items include "I have been so unhappy that I have been crying" and "I have blamed myself unnecessarily when things went wrong". A cut off score of ≥13 is recommended when screening for major depression (Murray and Cox, 1990). Cut off scores were used to determine the prevalence of probable major depression. Continuous scores were used for all other analyses. The EPDS is well validated within postnatal and non-postnatal populations, as well as in Australian populations (Cox et al., 1996;Murray and Cox, 1990), and has demonstrated good sensitivity and specificity in identifying women experiencing major depression (Levis et al., 2020). The Cronbach's Alpha (α) for the current sample at 3 months, 1, 4, and 10 years postpartum was 0.87, 0.88, 0.90, and 0.88, respectively.

Stressful life events and social health issues
Stressful life events and social health issues were assessed at 4 and 10 years postpartum using 20 items drawn from the Pregnancy Risk Assessment and Monitoring System (PRAMS) (Gilbert et al., 1999). Women were asked to report the occurrence of a range of stressful life experience across the previous 12 months. Examples included (a) major injury or illness, (b)started a new relationship, (c) death of a family member, (e) lost job when wanted to work, (f) you or your partner had trouble with alcohol or drugs. Women's reports on each item were summed to create a total stressful life events and social health issues score.

Data analysis
Analyses were conducted in two steps as outlined below. Descriptive statistics for the sample characteristics and study variables were computed using SPSS Version 27. In Step 1 Latent Class Analysis (LCA) of IPV experiences across timepoints was used to determine patterns of women's experiences of physical and emotional IPV across the first 10 years of motherhood. This approach was deemed appropriate given the IPV indicators were dichotomous variables.
Step 2 involved estimating a latent growth model to describe the course of depressive symptoms across the first 10 years of motherhood for the overall sample. Multiple group analysis was then applied to this latent growth model to estimate the trajectories of depressive symptoms for each IPV class identified at Step 1. The latent class and latent growth analyses described below were conducted using Mplus Version 8, and missing data in all models were handled using full information maximum likelihood (Muthén and Muthén, 2010).

Step 1: determining trajectories of women's experience of physical and emotional IPV across the first 10 years of motherhood
Latent Class Analysis (LCA) was conducted to identify groups of women based on their report of emotional and physical IPV on the CAS (0= no; 1= yes) in the first year, fourth year, and tenth year postpartum. This method involved identifying the smallest number of classes starting with a 1-class model and fitting successive models with increasing numbers of classes. The model fit indexes of Likelihood ratio statistic (L 2 ), Bayesian Information Criterion (BIC), and Akaike Information Criterion (AIC) were used to compare model solutions, with lower statistics indicating better fitting models. Entropy provides an index for assessing the precision of assigning latent class membership, with higher probability values indicating greater precision of classification. Posterior probabilities were obtained for each case to indicate the probability of belonging to each class. The Vuong-Lo-Mendell-Rubin likelihood ratio test was also used to test for significant differences between the models. Finally, the size of the latent classes was also considered in identifying the best model solution, with group sizes of >20 preferred to enable examination of classes within subsequent comparisons. Participants with data on experiences of physical and emotional IPV on the CAS at one or more timepoints were included in the analysis. One-way ANOVAs and Chi-square tests were conducted in SPSS Version 27 to determine (IBM Corp., 2021) whether the classes significantly differed on key demographic and social health issues including stressful life events and social health issues at 4 and 10 years postpartum.

Step 2: estimating the trajectories of women's depressive symptoms across the first 10 years of motherhood for each IPV trajectory
Trajectories of women's depressive symptoms across four postpartum time points (3 months, 1 year, 4 years and 10 years) assessed by the EPDS were identified using a latent growth model. Baseline characteristics of maternal age, highest educational qualification, participation in paid employment and language spoken at home were included as covariates as these demographic factors have been shown to relate to IPV and/or maternal depressive symptoms (FitzPatrick et al., 2022b;Foo et al., 2018;Woolhouse et al., 2012). This involved creating latent factors to represent the initial or baseline levels of the depressive symptom variable (i.e., the intercept) and the trajectory or change in that variable over time (i.e., the slope). The intercept factor was created with a fixed loading of 1.0 at each time point, whereas the slope factor was created with fixed values to each time point. The model was estimated using robust maximum likelihood estimation and assessed using the chi-square test and other practical fit indices including Tucker-Lewis index (TLI), the comparative fit index (CFI), and root-mean-square error of approximation (RMSEA). Indices for the TLI and CFI should exceed 0.90 for an acceptable fit (Bentler, 1990), and values close to or below 0.05 for the RMSEA were considered acceptable (Hu and Bentler, 1999).
Next, multiple group analysis was applied to this model to describe the course of depressive symptoms separately for women assigned to each of the IPV classes identified in Step 1. The latent growth parameter estimates were free to vary by IPV class. Participants assigned a class at Step 1 and with data on the EPDS at one or more time points were included in the analysis.

Sample characteristics
Of the 1507 nulliparous women who enroled in the study, 122 were excluded from analyses because they did not complete the CAS at any of the 1 year, 4 years, or 10 years timepoints. Demographic characteristics at baseline of the final sample of 1385 women are presented in Table 1. Most women included in the study were born in Australia, partnered and in paid employment at the time of study enrolment 3.2. Exposure to IPV and depressive symptoms across the study period Table 2 presents the proportion of women experiencing emotional and physical IPV in the first, fourth and tenth year postpartum. The means and standard deviations on the EPDS at 3 months, 1 years, 4 years and 10 years postpartum are presented in Table 2. The proportion of women scoring at or above the clinical cut point for maternal depressive symptoms was 6.8%, 7.8%, 9.3% and 9.9% at 3 months, 1, 4, and 10 years postpartum, respectively.

Latent class analysis to identify groups of women's experience of IPV over time
Latent class models specifying 1-6 models were estimated and are presented in Table 3. The AIC and BIC model indices were lowest within the 4-class model, whereas the Likelihood ratio statistic (L 2 ) continued to decrease as size class increased. The Vuong-Lo-Mendell-Rubin likelihood test indicated a significant difference between the 3-and 4-class models, suggesting that the 4-class model gives significant improvement in fit over the 3-class model. There were no significant differences between the 4-and 5 class models or the 5-and 6-class models. The 5-and 6-class models also contained classes with very small group sizes (<20).
On the basis of this information, the 4-class model was accepted as the final model. The entropy value was high, and the classification probabilities for latent membership were high (Class 1: 0.95, Class 2: 0.69, Class 3: 0.90, Class 4: 0.86) suggesting acceptable precision in assigning Note. Non-missing data varies between N = 1360-1385.

Table 2
Maternal exposure to IPV and depressive symptoms across the study period (N = 1385).  Table 4 and Fig. 1 present the probability of women in the identified classes experiencing emotional and physical IPV at 1 year, 4 years, and 10 years postpartum. Class 1 was the largest class, referred to as 'Minimal IPV' (n = 1097, 79.2%) as these women had very low probabilities of physical and emotional IPV across all time points. Class 2 referred to as 'Early exposure to IPV' comprised 150 women (10.8%) who had a moderate to high probability of physical and emotional IPV in the first and fourth years postpartum, whereas by 10 years postpartum, their probability of physical and emotional IPV had markedly decreased. Class 3 referred to as 'Increasing IPV' comprised 88 women (6.4%) whose probability of experiencing both physical and emotional IPV increased over time. By 10 years postpartum, their probability of IPV was 22% for physical and 100% for emotional IPV. Finally, Class 4 was the smallest class, referred to as 'Persistent IPV' (n = 50, 3.6%). Women in this class had moderate to high probability of physical and emotional IPV at all time points. Table 5 presents the sociodemographic and social health characteristics for each of the four IPV classes. Significant differences were noted in stressful life events and social health issues at 4 and 10 years postpartum across the four IPV classes. Significant differences were also noted across a range of demographic and social health issues. Most notably, differences were noted across employment status, Table 3 Model fit indices for latent class models for IPV from 1 year to 10 years postpartum. Note: L 2 = Likelihood-ratio statistic, BIC= Bayesian Information Criterion, AIC= Akaike Information Criterion.   Table 5 reports the parameter estimates for the latent growth model. Next, multiple group analysis was conducted to estimate the course of depressive symptoms (estimated means for each time point) for each IPV class identified at Step 1, whilst adjusting for the covariates. The model was an acceptable fit to the data,  Table 5 presents the parameter estimates for the latent growth factors for each IPV class. The estimated means of depressive symptoms at each timepoint for each IPV class are depicted in Fig. 2. Briefly, women in the 'Minimal IPV' class (class 1) had low levels of depressive symptoms across all timepoints. Women in the 'Early IPV' class (class 2) had higher depressive symptoms in the first year postpartum which decreased slightly over time. Women in the 'Increasing IPV' class (class 3) had stable depressive symptoms in the first year postpartum, with symptoms increasing at 4 years postpartum and again at 10 years postpartum. Finally, women in the 'Persistent IPV' class (class 4) had the highest levels of depressive symptoms within the first year postpartum, which steadily increased at 4 years and again by 10 years postpartum. Note. Denominator may vary because of missing data. 3.4 Trajectories of women's depressive symptoms across the first 10 years of motherhood for each IPV class. Table 5 Results of the latent growth modelling of depressive symptoms for the overall sample and 4-class model. This is one of the first studies to examine the course of mental health problems across specific patterns of emotional and physical IPV throughout the first decade of motherhood. We identified four unique classes of IPV experienced by women across the 10 years following the birth of their first child. The majority of women in our sample (79%) experienced minimal or low probability of IPV across the study period. The remaining classes were women who had experienced different patterns of exposure to physical and emotional IPV over time, including (a) those with exposure to IPV in the early years (1 and 4 years postpartum) but a decreased likelihood of exposure in the tenth year (10.8%), (b) those who experienced an increasing likelihood of IPV across time (6.4%), and (c) those who experienced persistent IPV across the study period (4%). These classes who experienced varying patterns of IPV across the study period, all experienced elevated trajectories of depressive symptoms compared to the minimal IPV class. Women who experienced a persistent pattern of IPV experienced the highest depressive symptoms at 12 months postpartum with these symptoms increasing in severity over time.
Our findings demonstrate how experiences of IPV change over time for some women, with a substantial proportion of our sample (17.2%) experiencing either an increase or decrease in IPV across the first decade of motherhood, with a smaller proportion (4%) experiencing persistent IPV across time. This finding points to changes occurring within families which have led to an increase, decrease, or in some cases cessation in violence. A smaller proportion (4%) experienced persistent IPV across time. This group also had the highest proportion of women who, at the time of study enrolment were recipients of a health care card, did not have a post high school qualification, and were not in a relationship. This is consistent with research linking IPV and socioeconomic disadvantage (Capaldi et al., 2012). It is possible that women within this IPV class face additional barriers to seeking support which may increase their vulnerability to ongoing or reoccurring IPV. Although past research has identified a range of risk factors associated with experiences of IPV (e.g., being younger, lower socioeconomic status, growing up in a home where there was violence) (Capaldi et al., 2012;Yakubovich et al., 2018), there remains limited evidence as to what factors may be associated with the changes in IPV trajectories suggested in our findings. This knowledge may be fundamental to understanding how to prevent and reduce IPV within our communities. Moreover, the presence of IPV in the lives of women across the first decade of motherhood underscores the need for health care providers to have a heightened sensitivity to the likelihood of IPV occurring within families. This is particularly important given the opportunity for intervention which exists within the early parenting period, with many families having increased health service use during this time.
This is one of few published studies to include both physical and emotional IPV in its investigation of IPV trajectories over time. Our study found that within each IPV pattern, the probability of experiencing emotional IPV was higher than that of physical IPV across most time points, but that the probability of physical IPV largely mirrored that of emotional IPV. This is consistent with previous research that suggests that physical violence often occurs in the presence of other types of abuse (FitzPatrick et al., 2022b;Hegarty et al., 2013). This may be particularly likely for the Persistent IPV class where the probabilities of experiencing emotional and physical IPV were highest, and most similar to each other. Across the other IPV classes there were some subtle differences worth noting. For example, women within the Early IPV exposure group experienced a decreased risk of both physical and emotional IPV over time. It is likely that some women within this group experienced a reduction or cessation of physical abuse over time but continued to experience emotional IPV. Although it is established that emotional abuse may occur in the absence of other forms of IPV (Coker et al., 2000a;FitzPatrick et al., 2022b;Mendonça and Ludermir, 2017), there has been limited previous evidence documenting this specific A. Fogarty et al. pattern.
The analysis of mental health trajectories demonstrated that women with a pattern of IPV at any time point experience elevated depression symptoms, underscoring the mental health burden associated with IPV (Bacchus et al., 2018;Brown et al., 2021;Devries et al., 2013). The trajectory of depressive symptoms experienced by the Early IPV class demonstrated a small decrease in symptoms as IPV risk also decreased, indicating some improvement in mental health in the absence of IPV. However, these women continued to experience elevated depressive symptoms, suggesting a sustained impact on mental health in the absence of ongoing IPV. This finding is consistent with earlier research indicating experiences of IPV can continue to impact mental health following cessation of abuse (Campbell and Soeken, 1999;Chuang et al., 2012;Ford-Gilboe et al., 2009). However, other studies report different findings (Dutton et al., 2004), including an Australian study that recruited participants via general practice clinics. This study by Hegarty et al. (2022) found that women who reported IPV at baseline but not at a 4 year follow up were no more likely to experience depressive symptoms at 4 year follow-up than women who reported minimal or no IPV across timepoints . As all participants in this study were recruited via general practice clinics, it may be that women within this study received some form of mental health support, contributing to their symptoms resolving over time. Our finding that depressive symptoms persist in the absence of IPV highlights the importance of available and accessible psychological support and pathways for healing for women who have experienced IPV, regardless of proximity since violence.
The Persistent IPV class experienced the highest prevalence of depressive symptoms at all timepoints within the study, emphasising the considerable impact of chronic exposure to IPV on mental health (Brown et al., 2020;Lagdon et al., 2014). The depressive symptoms of women within the Increasing IPV class followed a very similar trajectory, steadily increasing across the study period. Notably, women within this Increasing IPV class had a considerably lower probability of experiencing physical IPV across all timepoints compared to women in the Persistent IPV class. Despite this, depressive symptoms for both classes were at a similar point at the 10-year follow-up. This finding provides further evidence to the growing body of research documenting the impact of emotional IPV on women's mental health (Coker et al., 2000b;Dokkedahl et al., 2022;Lagdon et al., 2014). Given the high prevalence of emotional IPV, continued efforts by governments, health services and communities to increase awareness around emotional IPV and its impact are essential.

Study strengths and limitations
This is one of the first prospective cohort studies to identify longitudinal trajectories of physical and emotional IPV experienced by women in the first decade of motherhood and the course of depressive symptoms across the same period. Our study assessed both physical and emotional IPV enabling identification of nuanced patterns of IPV across the study period. Finally, our study draws on a community-based sample of first-time mothers and is therefore inclusive of women who did and did not report experiences of IPV and poor mental health to primary care and/or mental health services.
Despite these strengths, the study has several limitations. Our measure of IPV assessed the occurrence of violence in the 12 months prior to each follow-up (1 year, 4 years, and 10 years postpartum) and we were not able to account for experiences of IPV outside of these times. Moreover, our measure of IPV was not designed to assess experiences of sexual IPV, and therefore we were unable to include consideration of exposure to sexual IPV within our analysis. This is a considerable limitation as it is possible that different or more nuanced patterns of IPV might have been identified with the availability of sexual IPV data and data at other time points not captured within the current study. Despite this, previous research indicates that sexual IPV often occurs concurrently with emotional and/or physical IPV (Coker et al., 2000a;Smith et al., 2002). Therefore, it is likely that the majority of women experiencing sexual IPV within our sample were captured through their reporting of physical and/or emotional IPV. Our analyses were also not able to consider whether changes in occurrence of IPV occurred a result of leaving a relationship or beginning a new one. Future research could aim to further explore this, along with other factors (such as formal and informal support, employment, etc.) that may be associated with a decrease or increase in IPV. Furthermore, our study focused on women's experiences of depressive symptoms over time. The long-term course of other mental health problems associated with IPV including anxiety and posttraumatic stress symptoms are also important to investigate. It is important to note that our sample comprised women during the first decade of motherhood. It is possible that patterns of IPV differ between women with and without children given the evidence that decisions to stay in or leave relationships are often dictated based on the physical and emotional safety of children (Fogarty et al., 2019;Rhodes et al., 2010;Sani and Pereira, 2020). Moreover, it is also important to note other factors which may have impacted women's depressive symptoms during this period. For example, our findings demonstrated that there were significant differences observed between the IPV classes on stressful events and social health issues. Our findings should be interpreted within consideration of the broader context of women's lives in mind. Lastly, the current paper focused solely on the impact of IPV patterns on the mental health of women, however there is an urgent need to understand how exposure to different pattens of IPV over time also impact the mental health of children.

Implications and conclusions
Our findings highlight important implications for research, policy, and clinical practice. Firstly, despite advances in IPV research across the last decade, there remains a need to generate evidence regarding opportunities within families contributing to a cessation in IPV. This knowledge is essential to inform violence prevention strategies. Second, many women within our study experienced an escalation of risk of IPV across a 10-year period. Early identification and intervention efforts have the potential to identify IPV prior to this escalation occurring. The early parenting period represents an opportune time for identification of IPV and mental health problems within families, due to an increase in health service use for most families (Gunn et al., 1996). For identification efforts to be effective, appropriate referral avenues must be accessible and systemic barriers to identification including clinician training, time constraints and privacy addressed Hudspeth et al., 2022).
Our findings also highlight the enduring mental health symptoms faced by women who experience IPV. This was apparent for all IPV classes including those with relatively low likelihood of physical IPV in comparison to emotional IPV. Moreover, the continued experience of depressive symptoms for those in the Early Exposure IPV class highlights that cessation of IPV is not enough to protect mental health. There is an urgent need for an increased availability of trauma-informed mental health support which aims to identify those experiencing abuse and promote healing and recovery once safety is established. Responses to IPV must include access to services to treat mental health symptoms arising from prolonged experiences of trauma and a greater recognition of the prevalence and impact of emotional IPV. Such responses are important to support the healing and recovery of women and their children following exposure to violence.

Funding sources
The Mothers' and Young People's Study (Formerly the Maternal Health Study) was supported by project grants from the Australian National Health and Medical Research Council (NHMRC) (Nos. 199222, 433006, 491205 and 2000842)

Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.