Association Between Intimate Partner Violence Against Women and Mental Health: A Cross-Sectional Study in Greece

1. Introduction

Intimate Partner Violence (IPV) against women, one of the most common forms of violence worldwide [1], constitutes a significant public health problem and a fundamental violation of human rights [2]. The World Health Organization (WHO) and the United Nations General Assembly define IPV as any behavior by a current or former partner that leads to physical, sexual, psychological harm to a woman, including threats, controlling behaviors and coercion [2,3]. Aiming for the complete elimination of violence against women, early global efforts were formalized in 1993 when the United Nations General Assembly adopted the “Declaration on the Elimination of Violence against Women”, providing a vital framework for both national and global action [4,5]. However, more than three decades later, the prevalence and impact of IPV remain alarmingly profound. Recent WHO estimates (2023) highlight the magnitude of this crisis, revealing that nearly one-third of women and girls globally (about 840 million) experience physical and/or sexual IPV or non-partner sexual violence at some point in their lives [6]. Specifically, regarding IPV, the data show that a quarter of all women who have ever been in a relationship (25.8% of those aged 15-49 years and 24.7% of those 15 years and older) have faced physical and/or sexual violence from a current or former male intimate partner at least once in their lifetime [6].

In the most severe cases, the consequences of this violence escalate to extreme fatal outcomes. In 2024, it is estimated that around 50,000 women and girls were murdered by their intimate partners or other family members, equating to a global average of 137 femicides per day, or one murder every 10 minutes [7]. Within Europe, findings from the gender-based violence survey indicate that 17.7% of women in the EU-27 have been subjected to physical violence or threats and/or sexual violence by an intimate partner during their lifetime, a figure that rises to 31.8% when psychological violence is included [8].

Mirroring these broader global and European trends, the landscape of IPV in Greece is equally alarming [9]. According to European gender-based violence survey, an estimated 41.8% of ever partnered Greek women (aged 18-74) have experienced physical, sexual or psychological violence from an intimate partner at some point in their lives [8]. The onset of the COVID-19 pandemic and the subsequent restrictive measures acted as a major catalyst for IPV increase [10,11]. Strict lockdown measures trapped women in environments with their abusers, restricting their access to support networks. Highlighting this impact, a comprehensive European Parliament survey revealed that 93% of Greek women agreed that the pandemic increased physical and emotional violence against women in their country compared to an average of 77% across other European nations [10].

Moreover, Greek official police data document a consistent increase in reports of domestic violence against women over the past decade [9]. Most notably, female victims nearly doubled from 9,886 in 2023 to 18,640 in 2024, followed by a further 3% rise in 2025 (19,167) [9]. The severity of these ongoing incidents is underscored by the 349 “panic button” application activations (from March 2023 to May 2024), representing instances where police intervention was required [9]. The ultimate consequence of this ongoing crisis in Greece remains a critical concern, with 205 femicides documented from 2010 to 2025, nearly half of which (99) have taken place since 2020 [9].

Beyond physical injuries and fatal consequences, IPV contributes substantially to psychological morbidity [1]. Recent evidence demonstrates strong associations between IPV and a broad spectrum of short- and long-term adverse mental health outcomes. The most frequently identified consequences related to IPV include post-traumatic stress disorder (PTSD), depression, anxiety, panic disorder, perinatal disorders, suicidal ideation, sleep disturbances and substance abuse [1,12,13,14,15,16,17,18,19,20,21]. In addition, IPV exposure has been associated with more complex psychiatric conditions, such as bipolar disorder, schizophrenia, personality disorders and eating disorders [1,18]. Furthermore, prolonged exposure to IPV can severely worsen mental health conditions by intensifying fear, social isolation, and overall emotional dysregulation [12]. Victims frequently internalize their traumatic experiences, leading to self-blame, stigma, reduced self-esteem and life satisfaction, while creating severe difficulties in forming future relationships [12,15,19]. Crucially, evidence suggests that victims who experience many forms of IPV simultaneously, report significantly worse mental health outcomes than those who have been subjected to a single type of IPV [20].

Although the association between IPV and women’s mental health has been extensively investigated internationally, the extent to which these findings apply to the Greek population remains unclear [13,14,18]. Sociocultural norms, family structures, gender roles, economic conditions, access to support services, and help-seeking behaviors differ across countries and may influence both the experience of IPV and its psychological impact. Consequently, country-specific evidence is necessary to inform prevention strategies, healthcare interventions, and policy development. To date, no study has specifically examined the association between IPV against women and mental health outcomes in Greece. To the best of our knowledge, the present study is the first to address this important research gap by investigating the association between IPV and mental health among women in Greece. Ultimately, we hope that our findings may inform public health policy, clinical practice, and future research on IPV against women in Greece.

2. Materials and Methods

2.1. Study Design

A cross-sectional study was conducted in Greece using an online survey administered in July 2026. Data collection was performed using a structured questionnaire developed in Google Forms, which was distributed through social media channels, such as Facebook, Instagram, TikTok, and LinkedIn. The target population comprised women who had access to the internet and actively used Facebook, Instagram, TikTok or LinkedIn. The survey link was posted on publicly accessible pages, and disseminated through personal and professional networks of the scholars. To enhance visibility and participation, recruitment posts were reshared periodically over a four-day period. No financial or other incentives were provided to participants. Given the open accessibility of the survey link and the inability to determine the number of women who viewed the recruitment posts, an accurate response rate could not be calculated. Participation was entirely voluntary, with women self-selecting into the study by accessing the survey link and completing the questionnaire. Eligibility criteria included being a woman, having an intimate relationship at least for the last 12 months, and providing informed consent prior to participation. After all, women were recruited through a non-probability convenience sampling strategy. Thus, a convenience sample was obtained. The study was conducted and reported in accordance with the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines [22].

To reduce selection bias, the participant information sheet described the study as an investigation of intimate relationships rather than explicitly focusing on abusive behavior toward women. This approach was adopted to encourage participation from women regardless of their experiences of IPV, thereby reducing the likelihood of disproportionate recruitment of women with a history of victimization. Consequently, the potential for overestimating the prevalence of IPV within the study sample was reduced.

Sample size estimation was conducted using G*Power software (version 3.1.9.2). Based on a multivariable analytical model comprising 10 variables (four independent predictors and six confounding variables), the minimum required sample size was calculated assuming a conservative small effect size (f² = 0.02), a significance level (α) of 0.05, and a statistical power of 95%. Under these assumptions, a minimum sample of 921 women was required. The selection of a conservative effect size was intended to ensure sufficient statistical power to detect small yet potentially meaningful associations between the predictor variables and the study outcomes.

2.2. Measurements

Data on women demographic characteristics were collected, including age (continuous variable), marital status (singles/married/living with partner/divorced/widows), children (continuous variable), educational level (elementary school/high school/college degree), and employment status (employees or not). Also, we measured economic status with a self-reported scale ranging from 0 (very poor) to 10 (excellent).

We used the Woman Abuse Screening Tool (WAST) to measure levels of IPV against women [23,24]. The WAST enables the detection of physical, psychological/emotional, and sexual violence among women in intimate relationships, allowing for early intervention and referral to specialized support and assistance services. Several systematic reviews support the use of the WAST as a valid proxy for IPV in different healthcare settings such as primary healthcare settings, hospital-based settings, and emergency departments [25,26,27,28]. These reviews found that the WAST has strong psychometric properties, and, thus, it could be used as a reliable and valid screening tool to identify women experiencing past-year IPV. For instance, the largest study (n=5605) evaluated the WAST and reported a sensitivity of 87% (95% confidence interval = 85% to 90%) and specificity of 89% (95% confidence interval = 88% to 90%) in a population of women from a variety of clinical settings [29]. Additionally, a great advantage of the WAST is that assesses all areas of IPV (i.e., physical, psychological/emotional, and sexual violence), and, thus, could be used in different clinical settings and populations. For instance, the U.S. Preventive Services Task Force recommends the WAST as a reliable and valid screening tool to assess IPV among women [30]. After all, the WAST is considered to be a valid proxy of IPV and can be used as an IPV screening tool. The WAST includes eight items that refer to intimate relationships during the last year. Answers are on a three-point Likert scale; never (0), sometimes (1), often (2). Total score ranges from 0 to 16. The higher the total score, the greater the IPV. Based on the cut-off value proposed by the developers of the WAST, women scoring ≥ 5 on the instrument are considered likely to be victims of IPV, while those scoring < 5 are classified as unlikely to be experiencing violence. We used the valid Greek version of the WAST [31]. The Cronbach’s alpha for the WAST was 0.778 in our sample indicating good reliability.

The lifetime prevalence of physical, psychological/emotional, and sexual IPV was assessed using three dichotomous (yes/no) self-reported items. Specifically, women were asked whether they had ever experienced: (a) physical abuse perpetrated by an intimate partner, including behaviors such as slapping, punching, kicking, pushing, biting, hair pulling, assault with objects or weapons, confinement, restraint, strangulation, or suffocation; (b) psychological/emotional abuse, including ridicule, persistent criticism, humiliation, threats, blackmail, bullying, gaslighting or excessive control over personal activities, finances, social relationships, or appearance; and (c) sexual abuse, including any form of non-consensual sexual contact, rape, sexual harassment or sexting. Responses were recorded as “yes” or “no” for each type of IPV. Percentage of positive answers in the total sample was considered as the lifetime prevalence of physical, psychological/emotional, and sexual IPV.

Anxiety-like and depressive-like symptoms were assessed using the Patient Health Questionnaire-4 (PHQ-4) [32]. The PHQ-4 is a brief self-report screening instrument consisting of four items, including two items measuring anxiety-like symptoms and two items measuring depressive-like symptoms. It should be emphasized that the instrument evaluates symptoms indicative of anxiety and depression rather than establishing clinical diagnoses or measuring actual levels of anxiety and depressive disorders. Responses are rated on a four-point Likert scale ranging from 0 (not at all) to 3 (nearly every day). Scores for both the anxiety-like symptoms and depressive-like symptoms subscales range from 0 to 6, with higher scores reflecting greater symptom severity. Developers of the PHQ-4 suggest that individuals with scores of 3 or higher should be classified as depressed or anxious. For the purposes of this study, the validated Greek version of the PHQ-4 was employed [33]. Cronbach’s alpha coefficients were 0.813 for the anxiety-like symptoms subscale and 0.768 for the depressive-like symptoms subscale, indicating good reliability within the study sample.

Sleep quality was assessed using the single-item Sleep Quality Scale (SQS) [34]. The instrument evaluates overall perceived sleep quality through a single question: “During the past 7 days, how would you rate your sleep quality overall?” Participants are instructed to consider multiple aspects of their sleep experience, including sleep duration, ease of falling asleep, frequency of early awakenings, and the restorative quality of sleep. Responses are recorded on an 11-point scale ranging from 0 (worst sleep quality) to 10 (excellent sleep quality), with higher scores indicating better overall sleep quality. We used the valid Greek version of the scale [35].

Self-esteem was assessed using the Single-Item Self-Esteem Scale (SISE). The SISE is developed as a brief alternative to the Rosenberg Self-Esteem Scale [36]. The SISE is a single-item measure designed to evaluate global self-esteem. Participants respond to the item “I have high self-esteem” using a 7-point Likert scale ranging from 1 (not very true of me) to 7 (very true of me), with higher scores indicating higher levels of self-esteem. Despite its brevity, the SISE has demonstrated strong convergent validity with the Rosenberg Self-Esteem Scale and has shown comparable predictive validity across a variety of psychological outcomes. Consequently, it represents a reliable and valid instrument for assessing global self-esteem in research settings [36].

Loneliness was assessed using the UCLA 3-Item Loneliness Scale (UCLA-LS-3) [37]. The instrument consists of three items designed to evaluate subjective feelings of loneliness. Responses are rated on a three-point Likert scale ranging from 1 (hardly ever) to 3 (often). Total scores range from 3 to 9, with higher scores indicating greater levels of loneliness. The validated Greek version of the scale was employed [38]. The UCLA-LS-3 demonstrated very good psychometric properties in the study sample since the Cronbach’s alpha coefficient of 0.840.

Life satisfaction was assessed using the Satisfaction with Life Scale (SWLS), a widely used five-item self-report instrument designed to measure individuals’ global cognitive evaluations of their overall life satisfaction [39]. The SWLS assesses subjective well-being based on respondents’ personal standards and judgments rather than objective life circumstances or externally defined criteria. Participants rate each of the five items on a 7-point Likert scale ranging from 1 (strongly disagree) to 7 (strongly agree). Total scores range from 5 to 35, with higher scores indicating greater levels of life satisfaction. We used the valid Greek version of the scale [40]. The Cronbach’s alpha for the SWLS was 0.974 in our sample indicating very good reliability.

2.3. Ethical Issues

The study protocol received ethical approval from the Ethics Committee of the Faculty of Nursing, National and Kapodistrian University of Athens (approval number 82; June 17, 2026). The study was conducted in accordance with the ethical principles outlined in the Declaration of Helsinki [41]. Participation was voluntary and anonymous. Prior to enrolment, potential participants were provided with detailed information regarding the study objectives, procedures, voluntary nature of participation, and their right to withdraw from the study at any time without penalty. Electronic informed consent was obtained from all women before access to the questionnaire was granted, through the selection of an “I agree to participate” option. Data were collected using an anonymous online survey, and no personally identifiable information was obtained. To safeguard participant privacy and maintain confidentiality, the questionnaire was administered through a secure online platform with restricted access available only to authorized members of the research team. All study data were stored in password-protected files, and IP addresses or other indirect identifiers were not collected. Furthermore, responses were analyzed and reported exclusively in aggregate form to minimize any risk of participant identification and to ensure compliance with applicable data protection and ethical standards.

2.4. Statistical Analysis

Categorical variables are summarized using absolute (n) and relative (%) frequencies, whereas continuous variables were described using means, standard deviations (SD), medians, interquartile ranges, minimum values and maximum values. The distribution of continuous variables was assessed using the Kolmogorov-Smirnov test in conjunction with visual inspection of Q-Q plots, confirming normality. We used WAST score, lifetime physical IPV, lifetime psychological/emotional IPV, and lifetime sexual IPV as independent variables. We considered anxiety-like symptoms score, depression-like symptoms score, quality of sleep, self-esteem score, loneliness score, and life satisfaction score as the dependent variables. Initially, univariate associations were examined through univariate linear regression. Subsequently, multivariable linear regression models were constructed to estimate the independent effect of IPV against women on mental health outcomes, adjusting for potential confounding factors. We considered demographic characteristics of women as potential confounders. Results are reported as both unadjusted and adjusted unstandardized B coefficients with corresponding 95% confidence intervals (CIs) and p-values. Also, standardized coefficient betas are presented. Multicollinearity among independent variables was assessed through the Variance Inflation Factor (VIF), with values exceeding 5 indicating substantial multicollinearity. VIFs ranged from 1.422 to 2.046, and, thus, were within acceptable limits. Model assumptions were further evaluated by inspecting histograms of residuals for normality and scatterplots of residuals versus predicted values to assess homoscedasticity and linearity. Outlier diagnostics were also conducted. Outcome outliers were identified through studentized residuals, with absolute values greater than ±3 indicating potential outliers. Predictor outliers were assessed through leverage values, with a threshold of (2k + 2) / n, where k represents the number of independent variables (k = 10) and n the sample size (n = 1212), yielding a cut-off of 0.0182. Influential observations were evaluated through Cook’s distance, with values exceeding 1.0 indicating undue influence on model estimates. No influential observations or significant outliers were detected in the multivariable models (see Table 4, Table 5, Table 6, Table 7, Table 8 and Table 9 for details). Pearson’s correlation coefficients were calculated to evaluate correlations between normally distributed scale scores. Statistical significance was defined as a two-sided p-value < 0.05. All analyses were conducted using IBM SPSS Statistics for Windows, version 28.0 (IBM Corp., Armonk, NY, USA).

3. Results

3.1. Demographic Characteristics

The study sample comprised 1212 women. The mean age was 35.91 years (SD = 9.84), with a median age of 35 years (interquartile range = 16), a minimum age of 19 years, and a maximum age of 74 years. More than half of women were married or living with their partner (57.6%), while 35.5% were singles, 6.6% were divorced, and 0.3% were widows. In our sample, 42.2% of women had children. Also, 50.3% of women held a college degree, while 49.7% had completed secondary (high school) education. Most women (91.7%) have been working. The mean self-reported financial status score was 6.35 (SD = 1.57), with a median of 7 (interquartile range = 2), a minimum value of 0, and a maximum value of 10. Detailed demographic characteristics of women are presented in Table 1.

3.2. Study Scales

The mean WAST score was 4.15. In the study sample, 32.4% (n = 393) of women had a WAST score of ≥5, indicating exposure to IPV according to the established cut-off criterion.

Mean anxiety-like symptoms and depression-like symptoms score was 3.17, and 2.90, respectively. In our sample, 50.5% (n=612) of women scored 3 or higher on the anxiety scale and could be classified as anxious. Moreover, 43.6% (n=528) scored 3 or higher on the depression scale and could be classified as depressed.

Moreover, mean sleep quality score was 5.17, mean self-esteem score was 4.83, mean loneliness score was 5.18, and mean life satisfaction score was 21.44. A comprehensive summary of descriptive statistics for all study scales is presented in Table 2.

Table 3 presents the Pearson correlation coefficients among the study scales. Scores on WAST were significantly correlated with higher levels of anxiety-like symptoms (r = 0.512, p-value < 0.01), depression-like symptoms (r = 0.563, p-value < 0.01), and loneliness (r = 0.535, p-value < 0.01). Additionally, we found a negative correlation between scores on WAST and sleep quality (r = -0.535, p-value < 0.01), self-esteem (r = -0.214, p-value < 0.01), and life satisfaction (r = -0.564, p-value < 0.01). Overall, these findings suggest that greater exposure to current IPV was correlated with poorer psychological well-being, characterized by elevated levels of anxiety-like symptoms, depressive-like symptoms, and loneliness, as well as lower sleep quality, self-esteem, and life satisfaction.

3.3. Lifetime Intimate Partner Violence

Psychological/emotional IPV was the most commonly reported form of violence, followed by physical and sexual IPV. In particular, the lifetime prevalence of psychological/emotional IPV was 68.4% (n = 829), while the lifetime prevalence of physical IPV was 38.5% (n = 467). Lifetime sexual IPV was reported by 19.3% of women (n = 234). Moreover, the lifetime prevalence of physical and psychological/emotional IPV simultaneously was 37.9% (n=459), the lifetime prevalence of physical and sexual IPV simultaneously was 16.2% (n=196), the lifetime prevalence of psychological/emotional and sexual IPV simultaneously was 18.8% (n=228), and the lifetime prevalence of all types of IPV simultaneously was 18.8% (n=228).

3.4. Dependent Variable: Anxiety-like Symptoms

Univariate and multivariable linear regression analyses examining the association between women IPV and anxiety-like symptoms are presented in Table 4. In the adjusted model, higher levels of current women IPV (WAST score) were significantly associated with higher levels of anxiety-like symptoms (adjusted B coefficient = 0.287; 95% CI: 0.250 to 0.324; standardized beta = 0.514; p < 0.001). Additionally, levels of anxiety-like symptoms were higher among women that have experienced physical IPV (adjusted B coefficient = 0.311; 95% CI: 0.070 to 0.552; standardized beta = 0.080; p = 0.011), and sexual IPV (adjusted B coefficient = 0.568; 95% CI: 0.302 to 0.835; standardized beta = 0.136; p < 0.001) through their lives. Based on standardized coefficients, the association between current women IPV (WAST score) and anxiety-like symptoms was stronger than that observed for lifetime physical IPV and lifetime sexual IPV. The final multivariable model explained 31.9% of the variance in anxiety-like symptoms score, while demonstrating acceptable multicollinearity diagnostics (VIFs ranged from 1.422 to 2.046). Model assumptions were also assessed and satisfied. The normality of residuals was supported by visual inspection (Supplementary Figure S1). Furthermore, scatterplot of residuals versus predicted values confirmed the assumptions of homoscedasticity and linearity across the multivariable model (Supplementary Figure S2). There were no influential observations/outliers in our multivariable model (Table 4).

Table 4. Linear regression models with anxiety-like symptoms score (Patient Health Questionnaire-4) as the dependent variable (n=1212).

Independent variables	Univariate models			Multivariable modela
	Unadjusted unstandardized B coefficient	95% CI for B	P-value	Adjusted unstandardized B coefficient	95% CI for B	Standardized coefficient beta	P-value	VIF
WAST score	0.286	0.259 to 0.313	<0.001	0.287	0.250 to 0.324	0.514	<0.001	2.046
Lifetime physical IPV	1.366	1.160 to 1.571	<0.001	0.311	0.070 to 0.552	0.080	0.011	1.767
Lifetime psychological/emotional IPV	1.003	0.780 to 1.226	<0.001	-0.065	-0.297 to 0.166	-0.016	0.581	1.489
Lifetime sexual IPV	1.485	1.227 to 1.743	<0.001	0.568	0.302 to 0.835	0.136	<0.001	1.422

^a Model is adjusted for age, marital status, children, educational level, work status, and financial status; R² for the final multivariable model = 31.9%; p-value for ANOVA < 0.001; Studentized residuals ranged from -2.494 to 2.535; Leverage values ranged from 0.003 to 0.016; Cook’s distance ranged from 0.000 to 0.008. CI: confidence interval; IPV: intimate partner violence; VIF: Variance Inflation Factor; WAST: Woman Abuse Screening Tool.

Table 4. Linear regression models with anxiety-like symptoms score (Patient Health Questionnaire-4) as the dependent variable (n=1212).

Independent variables	Univariate models			Multivariable modela
	Unadjusted unstandardized B coefficient	95% CI for B	P-value	Adjusted unstandardized B coefficient	95% CI for B	Standardized coefficient beta	P-value	VIF
WAST score	0.286	0.259 to 0.313	<0.001	0.287	0.250 to 0.324	0.514	<0.001	2.046
Lifetime physical IPV	1.366	1.160 to 1.571	<0.001	0.311	0.070 to 0.552	0.080	0.011	1.767
Lifetime psychological/emotional IPV	1.003	0.780 to 1.226	<0.001	-0.065	-0.297 to 0.166	-0.016	0.581	1.489
Lifetime sexual IPV	1.485	1.227 to 1.743	<0.001	0.568	0.302 to 0.835	0.136	<0.001	1.422

^a Model is adjusted for age, marital status, children, educational level, work status, and financial status; R² for the final multivariable model = 31.9%; p-value for ANOVA < 0.001; Studentized residuals ranged from -2.494 to 2.535; Leverage values ranged from 0.003 to 0.016; Cook’s distance ranged from 0.000 to 0.008. CI: confidence interval; IPV: intimate partner violence; VIF: Variance Inflation Factor; WAST: Woman Abuse Screening Tool.

3.5. Dependent Variable: Depression-like Symptoms

Linear regression models examining the association between women exposure to IPV and depression-like symptoms are presented in Table 5. Multivariable analysis identified that higher levels of IPV, as measured by the WAST score, were significantly associated with increased depression symptomatology (adjusted B coefficient = 0.335; 95% CI: 0.297 to 0.374; standardized beta = 0.557; p < 0.001). Furthermore, depression levels were significantly higher among women who reported a lifetime history of physical IPV (adjusted B coefficient = 0.259; 95% CI: 0.009 to 0.509; standardized beta = 0.062; p = 0.042) and sexual IPV (adjusted B coefficient = 0.531; 95% CI: 0.255 to 0.808; standardized beta = 0.103; p < 0.001). Based on standardized coefficients, the association between current women IPV (WAST score) and depression-like symptoms was stronger than the corresponding associations observed for lifetime physical and sexual IPV. The final multivariable model explained 37.0% of the variance in depression-like symptoms and demonstrated acceptable multicollinearity diagnostics, with VIFs ranging from 1.422 to 2.046. Model assumptions were thoroughly evaluated and satisfied. The normality of residuals was confirmed through visual inspection (Supplementary Figure S3), while scatterplot of residuals versus predicted values supported the assumptions of homoscedasticity and linearity (Supplementary Figure S4). No influential observations or significant outliers were identified in the multivariable model (Table 5).

Table 5. Linear regression models with depression-like symptoms score (Patient Health Questionnaire-4) as the dependent variable (n=1212).

Independent variables	Univariate models			Multivariable modela
	Unadjusted unstandardized B coefficient	95% CI for B	P-value	Adjusted unstandardized B coefficient	95% CI for B	Standardized coefficient beta	P-value	VIF
WAST score	0.339	0.311 to 0.367	<0.001	0.335	0.297 to 0.374	0.557	<0.001	2.046
Lifetime physical IPV	1.511	1.290 to 1.732	<0.001	0.259	0.009 to 0.509	0.062	0.042	1.767
Lifetime psychological/emotional IPV	1.246	1.008 to 1.484	<0.001	0.064	-0.177 to 0.304	0.014	0.604	1.489
Lifetime sexual IPV	1.601	1.324 to 1.879	<0.001	0.531	0.255 to 0.808	0.103	<0.001	1.422

^a Model are adjusted for age, marital status, children, educational level, work status, and financial status; R² for the final multivariable model = 37.0%; p-value for ANOVA < 0.001; Studentized residuals ranged from -2.049 to 2.558; Leverage values ranged from 0.003 to 0.016; Cook’s distance ranged from 0.000 to 0.009. CI: confidence interval; IPV: intimate partner violence; VIF: Variance Inflation Factor; WAST: Woman Abuse Screening Tool.

Table 5. Linear regression models with depression-like symptoms score (Patient Health Questionnaire-4) as the dependent variable (n=1212).

Independent variables	Univariate models			Multivariable modela
	Unadjusted unstandardized B coefficient	95% CI for B	P-value	Adjusted unstandardized B coefficient	95% CI for B	Standardized coefficient beta	P-value	VIF
WAST score	0.339	0.311 to 0.367	<0.001	0.335	0.297 to 0.374	0.557	<0.001	2.046
Lifetime physical IPV	1.511	1.290 to 1.732	<0.001	0.259	0.009 to 0.509	0.062	0.042	1.767
Lifetime psychological/emotional IPV	1.246	1.008 to 1.484	<0.001	0.064	-0.177 to 0.304	0.014	0.604	1.489
Lifetime sexual IPV	1.601	1.324 to 1.879	<0.001	0.531	0.255 to 0.808	0.103	<0.001	1.422

^a Model are adjusted for age, marital status, children, educational level, work status, and financial status; R² for the final multivariable model = 37.0%; p-value for ANOVA < 0.001; Studentized residuals ranged from -2.049 to 2.558; Leverage values ranged from 0.003 to 0.016; Cook’s distance ranged from 0.000 to 0.009. CI: confidence interval; IPV: intimate partner violence; VIF: Variance Inflation Factor; WAST: Woman Abuse Screening Tool.

3.6. Dependent Variable: Sleep Quality

Linear regression analyses examining the association between women exposure to IPV and sleep quality are summarized in Table 6. In the multivariable model, higher levels of IPV, as measured by the WAST score, were significantly associated with worse sleep quality (adjusted B coefficient = -0.425; 95% CI: -0.476 to -0.373; standardized beta = -0.517; p < 0.001). Additionally, women that have experienced lifetime physical IPV (adjusted B coefficient = -0.395; 95% CI: -0.730 to -0.059; standardized beta = -0.069; p = 0.021), lifetime psychological/emotional IPV (adjusted B coefficient = -0.402; 95% CI: -0.724 to -0.080; standardized beta = -0.067; p = 0.014), and lifetime sexual IPV (adjusted B coefficient = -1.156; 95% CI: -1.527 to -0.785; standardized beta = -0.163; p < 0.001) have also experienced worse sleep quality. Based on standardized coefficients, the association between current women IPV (WAST score) and sleep quality was stronger than the corresponding associations observed for lifetime physical, psychological/emotional and sexual IPV. The final multivariable model accounted for 39.2% of the variance in sleep quality and demonstrated acceptable multicollinearity (VIFs ranged from 1.422 to 2.046). Model assumptions were thoroughly evaluated and met. The normality of residuals was supported by visual inspection (Supplementary Figure S5), while scatterplot of residuals versus predicted values confirmed the assumptions of homoscedasticity and linearity (Supplementary Figure S6). No influential observations or significant outliers were detected in the multivariable model (Table 6).

Table 6. Linear regression models with sleep quality (Single-Item Sleep Quality Scale) as the dependent variable (n=1212).

Independent variables	Univariate models			Multivariable modela
	Unadjusted unstandardized B coefficient	95% CI for B	P-value	Adjusted unstandardized B coefficient	95% CI for B	Standardized coefficient beta	P-value	VIF
WAST score	-0.440	-0.479 to -0.400	<0.001	-0.425	-0.476 to -0.373	-0.517	<0.001	2.046
Lifetime physical IPV	-2.216	-2.515 to -1.918	<0.001	-0.395	-0.730 to -0.059	-0.069	0.021	1.767
Lifetime psychological/emotional IPV	-1.874	-2.196 to -1.552	<0.001	-0.402	-0.724 to -0.080	-0.067	0.014	1.489
Lifetime sexual IPV	-2.676	-3.045 to -2.306	<0.001	-1.156	-1.527 to -0.785	-0.163	<0.001	1.422

^a Model are adjusted for age, marital status, children, educational level, work status, and financial status; R² for the final multivariable model = 39.2%; p-value for ANOVA < 0.001; Studentized residuals ranged from -2.817 to 2.718; Leverage values ranged from 0.003 to 0.016; Cook’s distance ranged from 0.000 to 0.012. CI: confidence interval; IPV: intimate partner violence; VIF: Variance Inflation Factor; WAST: Woman Abuse Screening Tool.

Table 6. Linear regression models with sleep quality (Single-Item Sleep Quality Scale) as the dependent variable (n=1212).

Independent variables	Univariate models			Multivariable modela
	Unadjusted unstandardized B coefficient	95% CI for B	P-value	Adjusted unstandardized B coefficient	95% CI for B	Standardized coefficient beta	P-value	VIF
WAST score	-0.440	-0.479 to -0.400	<0.001	-0.425	-0.476 to -0.373	-0.517	<0.001	2.046
Lifetime physical IPV	-2.216	-2.515 to -1.918	<0.001	-0.395	-0.730 to -0.059	-0.069	0.021	1.767
Lifetime psychological/emotional IPV	-1.874	-2.196 to -1.552	<0.001	-0.402	-0.724 to -0.080	-0.067	0.014	1.489
Lifetime sexual IPV	-2.676	-3.045 to -2.306	<0.001	-1.156	-1.527 to -0.785	-0.163	<0.001	1.422

^a Model are adjusted for age, marital status, children, educational level, work status, and financial status; R² for the final multivariable model = 39.2%; p-value for ANOVA < 0.001; Studentized residuals ranged from -2.817 to 2.718; Leverage values ranged from 0.003 to 0.016; Cook’s distance ranged from 0.000 to 0.012. CI: confidence interval; IPV: intimate partner violence; VIF: Variance Inflation Factor; WAST: Woman Abuse Screening Tool.

3.7. Dependent Variable: Self-Esteem

Table 7 shows the results from the linear regression analyses with self-esteem as the dependent variable. The final multivariable model found a negative association between levels of IPV, as measured by the WAST score, (adjusted B coefficient = -0.060; 95% CI: -0.091 to -0.029; standardized beta = -0.147; p < 0.001) and self-esteem. Furthermore, women with a reported lifetime history of psychological/emotional IPV (adjusted B coefficient = -0.208; 95% CI: -0.401 to -0.016; standardized beta = -0.067; p = 0.034), also demonstrated significantly lower self-esteem. Standardized coefficients showed that the association between current women abuse (WAST score) and self-esteem was stronger than the association observed for psychological/emotional IPV. The final multivariable model explained 11.8% of the variance in self-esteem and demonstrated absence of multicollinearity, normality of residuals, homoscedasticity and linearity (Supplementary Figures S7 and S8). No influential observations or significant outliers were identified in the multivariable model (Table 7).

Table 7. Linear regression models with self-esteem (Single-Item Self-Esteem Scale) as the dependent variable (n=1212).

Independent variables	Univariate models			Multivariable modela
	Unadjusted unstandardized B coefficient	95% CI for B	P-value	Adjusted unstandardized B coefficient	95% CI for B	Standardized coefficient beta	P-value	VIF
WAST score	-0.087	-0.110 to -0.065	<0.001	-0.060	-0.091 to -0.029	-0.147	<0.001	2.046
Lifetime physical IPV	-0.299	-0.458 to -0.139	<0.001	-0.009	-0.210 to 0.191	-0.003	0.926	1.767
Lifetime psychological/emotional IPV	-0.430	-0.596 to -0.264	<0.001	-0.208	-0.401 to -0.016	-0.067	0.034	1.489
Lifetime sexual IPV	-0.303	-0.500 to -0.106	0.003	0.119	-0.103 to 0.341	0.034	0.292	1.422

^a Model are adjusted for age, marital status, children, educational level, work status, and financial status; R² for the final multivariable model = 11.8%; p-value for ANOVA < 0.001; Studentized residuals ranged from -2.719 to 1.927; Leverage values ranged from 0.003 to 0.016; Cook’s distance ranged from 0.000 to 0.011. CI: confidence interval; IPV: intimate partner violence; VIF: Variance Inflation Factor; WAST: Woman Abuse Screening Tool.

Table 7. Linear regression models with self-esteem (Single-Item Self-Esteem Scale) as the dependent variable (n=1212).

Independent variables	Univariate models			Multivariable modela
	Unadjusted unstandardized B coefficient	95% CI for B	P-value	Adjusted unstandardized B coefficient	95% CI for B	Standardized coefficient beta	P-value	VIF
WAST score	-0.087	-0.110 to -0.065	<0.001	-0.060	-0.091 to -0.029	-0.147	<0.001	2.046
Lifetime physical IPV	-0.299	-0.458 to -0.139	<0.001	-0.009	-0.210 to 0.191	-0.003	0.926	1.767
Lifetime psychological/emotional IPV	-0.430	-0.596 to -0.264	<0.001	-0.208	-0.401 to -0.016	-0.067	0.034	1.489
Lifetime sexual IPV	-0.303	-0.500 to -0.106	0.003	0.119	-0.103 to 0.341	0.034	0.292	1.422

^a Model are adjusted for age, marital status, children, educational level, work status, and financial status; R² for the final multivariable model = 11.8%; p-value for ANOVA < 0.001; Studentized residuals ranged from -2.719 to 1.927; Leverage values ranged from 0.003 to 0.016; Cook’s distance ranged from 0.000 to 0.011. CI: confidence interval; IPV: intimate partner violence; VIF: Variance Inflation Factor; WAST: Woman Abuse Screening Tool.

3.8. Dependent Variable: Loneliness

Table 8 shows the findings of the linear regression analyses with loneliness score, as measured by the UCLA 3-Item Loneliness Scale, as the dependent variable. In the final multivariable model, higher levels of current IPV (i.e., WAST score) were significantly associated with higher levels of loneliness (adjusted B coefficient = 0.219; 95% CI: 0.181 to 0.256; standardized beta = 0.385; p < 0.001). Additionally, women reporting a lifetime history of sexual IPV demonstrated significantly increased loneliness (adjusted B coefficient = 0.735; 95% CI: 0.467 to 1.002; standardized beta = 0.150; p < 0.001). Comparison of standardized coefficients indicated that the association between current women IPV (WAST score) and loneliness was stronger than that observed lifetime sexual IPV. The final multivariable model explained 33.8% of the variance in loneliness score and met all key regression assumptions, including the absence of multicollinearity, as well as normality of residuals, homoscedasticity, and linearity (Supplementary Figures S9 and S10). No influential observations or significant outliers were identified in the model (Table 8).

Table 8. Linear regression models with loneliness score (UCLA 3-Item Loneliness Scale) as the dependent variable (n=1212).

Independent variables	Univariate models			Multivariable modela
	Unadjusted unstandardized B coefficient	95% CI for B	P-value	Adjusted unstandardized B coefficient	95% CI for B	Standardized coefficient beta	P-value	VIF
WAST score	0.304	0.277 to 0.331	<0.001	0.219	0.181 to 0.256	0.385	<0.001	2.046
Lifetime physical IPV	0.950	0.733 to 1.167	<0.001	-0.116	-0.358 to 0.125	-0.029	0.345	1.767
Lifetime psychological/emotional IPV	0.986	0.759 to 1.213	<0.001	0.020	-0.213 to 0.252	0.005	0.869	1.489
Lifetime sexual IPV	1.410	1.146 to 1.674	<0.001	0.735	0.467 to 1.002	0.150	<0.001	1.422

^a Model are adjusted for age, marital status, children, educational level, work status, and financial status; R² for the final multivariable model = 33.8%; p-value for ANOVA < 0.001; Studentized residuals ranged from -2.499 to 2.828; Leverage values ranged from 0.003 to 0.016; Cook’s distance ranged from 0.000 to 0.014. CI: confidence interval; IPV: intimate partner violence; VIF: Variance Inflation Factor; WAST: Woman Abuse Screening Tool.

Table 8. Linear regression models with loneliness score (UCLA 3-Item Loneliness Scale) as the dependent variable (n=1212).

Independent variables	Univariate models			Multivariable modela
	Unadjusted unstandardized B coefficient	95% CI for B	P-value	Adjusted unstandardized B coefficient	95% CI for B	Standardized coefficient beta	P-value	VIF
WAST score	0.304	0.277 to 0.331	<0.001	0.219	0.181 to 0.256	0.385	<0.001	2.046
Lifetime physical IPV	0.950	0.733 to 1.167	<0.001	-0.116	-0.358 to 0.125	-0.029	0.345	1.767
Lifetime psychological/emotional IPV	0.986	0.759 to 1.213	<0.001	0.020	-0.213 to 0.252	0.005	0.869	1.489
Lifetime sexual IPV	1.410	1.146 to 1.674	<0.001	0.735	0.467 to 1.002	0.150	<0.001	1.422

^a Model are adjusted for age, marital status, children, educational level, work status, and financial status; R² for the final multivariable model = 33.8%; p-value for ANOVA < 0.001; Studentized residuals ranged from -2.499 to 2.828; Leverage values ranged from 0.003 to 0.016; Cook’s distance ranged from 0.000 to 0.014. CI: confidence interval; IPV: intimate partner violence; VIF: Variance Inflation Factor; WAST: Woman Abuse Screening Tool.

3.9. Dependent Variable: Life Satisfaction

Linear regression models with life satisfaction score as the dependent variable are shown in Table 9. After adjusting for confounders, we identified a negative association between women IPV (WAST score) and life satisfaction (adjusted B coefficient = -1.340; 95% CI: -1.466 to -1.215; standardized beta = -0.648; p < 0.001). Moreover, women reporting a lifetime history of sexual IPV exhibited significantly lower levels of life satisfaction (adjusted B coefficient = -1.319; 95% CI: -2.219 to -0.418; standardized beta = -0.074; p = 0.004). Standardized coefficients revealed that the association between WAST score and life satisfaction was stronger than that observed for lifetime sexual IPV. The final multivariable model accounted for 43.4% of the variance in life satisfaction score and satisfied all key regression assumptions, including the absence of multicollinearity, as well as normality of residuals, homoscedasticity, and linearity (Supplementary Figures S11 and S12). No influential observations or significant outliers were identified in the model (Table 9).

Table 9. Linear regression models with life satisfaction score (Satisfaction with Life Scale) as the dependent variable (n=1212).

Independent variables	Univariate models			Multivariable modela
	Unadjusted unstandardized B coefficient	95% CI for B	P-value	Adjusted unstandardized B coefficient	95% CI for B	Standardized coefficient beta	P-value	VIF
WAST score	-1.165	-1.262 to -1.069	<0.001	-1.340	-1.466 to -1.215	-0.648	<0.001	2.046
Lifetime physical IPV	-4.009	-4.792 to -3.227	<0.001	0.354	-0.460 to 1.169	0.025	0.393	1.767
Lifetime psychological/emotional IPV	-3.517	-4.346 to -2.688	<0.001	0.335	-0.447 to 1.117	0.022	0.401	1.489
Lifetime sexual IPV	-5.241	-6.201 to -4.282	<0.001	-1.319	-2.219 to -0.418	-0.074	0.004	1.422

^a Model are adjusted for age, marital status, children, educational level, work status, and financial status; R² for the final multivariable model = 43.4%; p-value for ANOVA < 0.001; Studentized residuals ranged from -2.812 to 2.764; Leverage values ranged from 0.003 to 0.016; Cook’s distance ranged from 0.000 to 0.010. CI: confidence interval; IPV: intimate partner violence; VIF: Variance Inflation Factor; WAST: Woman Abuse Screening Tool.

Table 9. Linear regression models with life satisfaction score (Satisfaction with Life Scale) as the dependent variable (n=1212).

Independent variables	Univariate models			Multivariable modela
	Unadjusted unstandardized B coefficient	95% CI for B	P-value	Adjusted unstandardized B coefficient	95% CI for B	Standardized coefficient beta	P-value	VIF
WAST score	-1.165	-1.262 to -1.069	<0.001	-1.340	-1.466 to -1.215	-0.648	<0.001	2.046
Lifetime physical IPV	-4.009	-4.792 to -3.227	<0.001	0.354	-0.460 to 1.169	0.025	0.393	1.767
Lifetime psychological/emotional IPV	-3.517	-4.346 to -2.688	<0.001	0.335	-0.447 to 1.117	0.022	0.401	1.489
Lifetime sexual IPV	-5.241	-6.201 to -4.282	<0.001	-1.319	-2.219 to -0.418	-0.074	0.004	1.422

^a Model are adjusted for age, marital status, children, educational level, work status, and financial status; R² for the final multivariable model = 43.4%; p-value for ANOVA < 0.001; Studentized residuals ranged from -2.812 to 2.764; Leverage values ranged from 0.003 to 0.016; Cook’s distance ranged from 0.000 to 0.010. CI: confidence interval; IPV: intimate partner violence; VIF: Variance Inflation Factor; WAST: Woman Abuse Screening Tool.

4. Discussion

To our knowledge, this is the first study to investigate the association between IPV and mental health outcomes among women in Greece. Our study sought to directly inform health policy by examining the association of women’s experiences of physical, sexual, and emotional IPV with depression, anxiety, sleep quality, self-esteem, loneliness and life satisfaction.

We observed high prevalence of lifetime IPV, with psychological/emotional IPV being the most common (68.4%), followed by physical (38.5%) and sexual IPV (19.3%). Furthermore, 32.4% of our women reported experiencing IPV in their current relationship. To understand these high prevalence rates, especially regarding psychological violence, sociocultural factors must be taken into account. While Greece has made notable institutional advancements in gender equality, such as being among the first countries to sign the Council of Europe’s Convention on preventing and combating violence against women in 2011 and establishing the Ministry of Social Cohesion and Family in 2023 [42], society continues to be heavily influenced by deep-rooted patriarchal structures and traditional gender roles. The tangible impact of these structures can be quantified by the European Institute for Gender Equality (EIGE), with Greece scoring 57 out of 100 and consistently ranking 22^nd out of the 27 EU Member States in the 2023 and 2024 editions of the Gender Equality Index [43]. Furthermore, the sociocultural environment frequently exposes IPV survivors to secondary victimization and stigma [44]. Literature indicates that negative public reactions, such as victim-blaming and shaming, directly contribute to the development of mental health problems, including depression and PTSD symptoms [44]. Consequently, many women internalize these societal judgements and accept psychological IPV as a normalized behavior, which significantly delays both the recognition of the abuse and subsequent help-seeking.

Crucially, based on earlier studies linking IPV with an increased risk of mental health consequences [13,14,18], our multivariate analyses demonstrated that higher overall levels of current IPV, as measured by the WAST, were consistently and independently associated with a pervasive deterioration across all the assessed domains of mental health and psychological well-being. Taken together, these particularly concerning estimates suggest that treating mental health problems in Greece also requires addressing the high rates of IPV and its profound impact on women’s health.

4.1. Anxiety-like Symptoms

Our findings suggested that current exposure to IPV (WAST score) was associated with increased anxiety-like symptoms. Additionally, a lifetime history of physical and sexual IPV independently predicted higher anxiety levels. Although anxiety typically functions as a natural physiological reaction to stress, it can evolve into a pathological and unmanageable condition [45]. Because women subjected to IPV are systematically exposed to constant fear and threats, this continuous psychological tension overwhelms their natural stress response, leading to unpleasant feelings and anxiety [46].

Our findings align with recent global literature across multiple contexts. Notably, a recent umbrella review included five systematic reviews and confirmed a strong association between IPV and anxiety disorders [18]. Specifically, White et al. (2024) reported that the odds of anxiety were highest following sexual (odds ratio [OR] = 2.34) and physical IPV (OR = 2.30), compared to psychological IPV (OR = 1.86) [14]. This is also supported by Spencer et al. (2023), who reported that any IPV exposure significantly increases the risk of anxiety disorders (relative risk = 2.57) [47]. Finally, Oram et al. (2014) provided additional evidence, noting that physical IPV is associated with a substantial increase in the risk of generalized anxiety (OR = 2.38) and panic disorder (OR = 1.9) [48].

4.2. Depression-like Symptoms

Our multivariable regression model identified current IPV (WAST score) as a significant predictor of depression. Moreover, we found a positive association between lifetime sexual IPV and lifetime physical IPV and depression-like symptoms. These findings are in accordance with a recent systematic review by Deacu et al. (2026), which noted that across the studies evaluating mental health outcomes of IPV female victims, the prevalence of depression ranged from 20% to over 50% [20]. Also, our findings align with recent systematic reviews which demonstrate that exposure to IPV is a major risk factor for subsequent depressive symptoms [13,14,16,18]. Specifically, Bacchus et al. (2018) observed a positive association between current IPV and depression (OR = 1.76) [13]. Building on this, a meta-analysis by Watson et al. (2025) reported that female IPV survivors face significantly increased odds of developing depression (OR = 1.92) [16]. Similarly, a meta-analysis by White et al. (2024) demonstrated that physical IPV had the strongest association with depression (OR = 3.14), followed by psychological IPV (OR = 2.54) and sexual IPV (OR = 2.04) [14]. The mechanisms underlying this strong association are complex, including both psychological and biological pathways [16,49]. IPV exposes victims to stressful experiences, often through behaviors that enforce social isolation and systematically diminish self-worth [49]. As a result, chronic exposure to trauma triggers severe psychological stress responses combined with endocrine and immune-inflammatory dysregulation, both of which are closely associated with depression [16].

4.3. Sleep Quality

Sleep disturbances are critical, yet frequently overlooked, physiological consequences of IPV. Our findings reveal a comprehensive deterioration of sleep quality among abused women. Current IPV and all types of lifetime IPV (physical, psychological/emotional and sexual) were significant predictors of poor sleep. Lifetime sexual IPV, in particular, demonstrated the most severe negative association. Recent evidence supports these results, indicating that IPV significantly increases the risk of impaired sleep [50]. For many victims, particularly those who have experienced sexual and physical IPV, the bedroom itself is often linked to traumatic events, acting as a potential trigger of psychological distress and fear that prevents good sleep quality [51]. Furthermore, chronic exposure to trauma fundamentally alters the reactivity of the stress response system, leading to hypervigilance and hyperarousal and thus disturbed sleep [51]. As outlined by Maharaj et al. (2022), sleep disturbances are frequently associated with IPV-induced PTSD or PTSD-like symptoms [51]. While nightmares and insomnia are the most widely recognized sleep disruptions associated with post-traumatic stress, evidence also suggests that a broader spectrum of physiological issues can be present, including sleep-disordered breathing, restless legs, and disruptive nocturnal behaviors [51]. Furthermore, it is shown that the physical trauma of IPV directly leads to sleep impairment. A recent study by Muller et al. (2025) demonstrated that women with IPV-related head injuries report significantly worse sleep quality [52].

4.4. Self-Esteem

The analysis of self-esteem provided valuable insights into the nature of psychological/emotional IPV. In the multivariable model, current IPV and lifetime psychological/emotional IPV was negatively associated with self-esteem. This finding aligns with the theoretical understanding of emotional IPV, which is fundamentally characterized by verbal and non-verbal abusing behaviors which intend to diminish the victim’s self-esteem and self-worth [15]. The mechanism driving this decline in self-esteem is deeply rooted in how victims process trauma. As highlighted by the American Psychiatric Association, survivors of domestic violence frequently internalize the continuous maltreatment, fostering an intense sense of self-blame and reduced self-esteem [12].

4.5. Loneliness

Women subjected to IPV frequently feel alone, despite being with others [53]. Our findings also support this reality, demonstrating that higher levels of current IPV are significantly associated with elevated loneliness among women. Additionally, women with a lifetime history of sexual IPV reported significantly higher levels of loneliness. Crucially, the development of loneliness is multifactorial. First, perpetrators actively employ controlling behaviors and psychological manipulation, such as gaslighting, to systematically isolate victims from their support networks [54]. However, this isolation is heavily compounded by the devastating reactions of the victim’s broader social environment. When survivors seek help and instead meet skepticism, silence and blame from personal networks and professional institutions, they often feel abandoned and silenced by the system that is meant to protect them [54]. This secondary victimization, an experience that survivors describe as even more painful than the violence of their abusers, significantly deepens their sense of loneliness [54].

4.6. Life Satisfaction

Our findings reveal a significant negative association between current abuse and women’s life satisfaction. Moreover, women reporting a lifetime history of sexual IPV exhibited significantly lower levels of life satisfaction. Given that life satisfaction serves as a core indicator of overall psychological health, these results underscore the severe impact of IPV on women’s well-being [55]. The mechanism driving this reduction is likely rooted in psychological distress associated with IPV. Specifically, as previously noted, exposure to IPV frequently results in anxiety, stress and depression [13,14,18], all of which severely compromise a victim’s overall well-being [56]. Consequently, IPV-related trauma triggers these adverse mental health conditions, which are significant predictors of low life satisfaction [57,58]. Furthermore, our data are highly consistent with recent evidence demonstrating that exposure to IPV is directly associated with decreased life satisfaction among women [15,59,60]. Specifically, Hui et al. (2021) found that women who have been subjected to physical IPV or unwanted sex from an intimate partner in the past 12 months had 2.05 higher odds of reporting poor life satisfaction [60].

4.7. Limitations

Several limitations of this study should be considered when interpreting the findings. First, the cross-sectional design precludes the establishment of temporal or causal relationships between IPV and mental health outcomes. Although significant associations were identified, it cannot be determined whether IPV contributed to poorer mental health, whether pre-existing mental health difficulties increase vulnerability to abusive relationships, or whether both are influenced by other factors. Second, the study employed a non-probability convenience sampling approach through online recruitment via social media platforms. As participation was voluntary and based on self-selection, the sample cannot be representative of all women in Greece. Women who use social media, have internet access, or have a particular interest in relationship-related issues may have been more likely to participate, potentially limiting the generalizability of the findings to the wider female population. Third, because the survey link was distributed through publicly accessible social media posts, it was not possible to determine how many individuals were exposed to the recruitment materials. Consequently, a response rate could not be calculated, making it difficult to assess the extent of potential non-response bias. Fourth, all variables were measured using self-report questionnaires, which are susceptible to information bias, including recall bias and social desirability bias. Participants may have underreported experiences of IPV due to stigma, fear, shame, or reluctance to disclose sensitive information. Similarly, self-reported mental health outcomes may not fully correspond to clinically diagnosed conditions. Fifth, IPV and mental health outcomes were assessed at a single point in time. The study therefore could not capture changes in IPV experiences or psychological well-being across the life course, nor could it examine the cumulative effects of IPV over time. Sixth, although validated screening instruments were employed, sleep quality and self-esteem were assessed using single-item measures. While these instruments have demonstrated acceptable psychometric properties, they may not capture the full complexity of constructs such as sleep quality and self-esteem as comprehensively as longer multidimensional assessments. Seventh, the online survey methodology may have excluded certain population groups, including women with limited digital literacy, restricted internet access, older age, lower socioeconomic status, or those living in geographically isolated areas. This may have introduced selection bias and further limited the representativeness of the sample. Eighth, despite efforts to minimize recruitment bias by describing the study as an investigation of intimate partner relationships rather than IPV, self-selection bias cannot be completely excluded. Women with either particularly positive or particularly negative relationship experiences may have been more motivated to participate than others. Finally, although several demographic characteristics were measured and controlled for during statistical analyses, residual confounding from unmeasured factors remains possible. For example, previous mental health history, childhood adversity, social support, substance use, personality characteristics, and exposure to other forms of violence were not assessed and may have influenced the observed associations. Despite these limitations, this study has important strengths, including the use of validated instruments to assess multiple dimensions of IPV and mental health, the inclusion of a relatively diverse sample of women from across Greece, and the generation of the first empirical evidence on the association between IPV and mental health outcomes among women in Greece. Consequently, the findings provide an important foundation for future longitudinal and nationally representative research in this field.

5. Conclusions

This study provides the first empirical evidence regarding the association between IPV and mental health among females in Greece. Our findings reveal an alarmingly high prevalence of both lifetime and current abuse, which is consistently and independently associated with severe psychological morbidity. Specifically, higher levels of current abuse, as well as a history of physical, psychological or sexual violence, significantly predict elevated anxiety and depressive symptomatology, increased loneliness, poor sleep quality, diminished self-esteem, and reduced overall life-satisfaction.

Consequently, addressing the mental health needs of women in Greece necessitates a comprehensive, patient-centered approach that directly targets gender-based violence. Policymakers and healthcare professionals must prioritize the integration of routine, safe screening for IPV (such as the application of the WAST) within primary care and psychiatric settings. Furthermore, it is important to establish culturally tailored psychological support services and implement broad public health campaigns aimed at eliminating stigma and victim-blaming. Future longitudinal and nationally representative studies are needed to further investigate the associations between IPV and mental health consequences, ultimately guiding the development of targeted, evidence-based interventions to protect, support and empower vulnerable women.