Associations of intimate partner violence with menstrual symptoms and healthcare-seeking among women in China: a nationwide cross-sectional survey

Intimate partner violence (IPV) constitutes a major global public health issue and a serious violation of women’s rights. In 2023, an estimated 608 million women aged 15 years and older worldwide had experienced IPV, including physical and/or sexual violence perpetrated by a current or former intimate partner. Concurrently, over 1 billion individuals aged 15 and above had experienced sexual violence in childhood. These experiences are closely associated with long-term adverse health outcomes, including depression and anxiety, chronic diseases and premature mortality. 1 IPV also remains a major barrier to achieving the United Nations Sustainable Development Goals (SDGs), particularly SDG 3, which aims to ensure healthy lives and promote well-being, and SDG 5, which aims to achieve gender equality and empower all women and girls. Accordingly, the WHO Global Plan of Action to End Violence against Women and Girls (2022–2030) highlights the need to address the health consequences of IPV, including reproductive health disparities. 2 3 The pathways linking IPV to poor health are complex and may involve chronic psychological stress, neuroendocrine dysregulation and sustained inflammatory responses, with potential consequences for multiple body systems, including reproductive health. 4 Menstrual symptoms, particularly primary dysmenorrhoea (PD) and premenstrual syndrome (PMS), are common gynaecological concerns among reproductive-aged women. These symptoms can impair quality of life, daily functioning, productivity and overall well-being. 5 Recent systematic reviews and meta-analyses have shown that dysmenorrhoea and premenstrual symptoms are highly prevalent across diverse settings, although estimates vary substantially according to diagnostic criteria, measurement tools and study populations. 6 7 Although biological factors are central to their occurrence, growing evidence suggests that psychosocial stress and adverse life experiences can exacerbate symptom severity and perception. 5 Given the intense psychosocial stress associated with IPV, IPV exposure may be an important but insufficiently examined factor related to more severe or frequent menstrual symptoms. However, existing epidemiological evidence on this association remains limited and inconsistent. Few studies have distinguished between IPV subtypes, including physical, sexual and psychological violence, or examined their specific associations with different menstrual symptoms. 7 – 9 Furthermore, the impact of IPV may extend beyond symptom generation to critically influence subsequent health-seeking behaviours. 10 11 Women exposed to IPV often face multiple barriers to healthcare, including fear, partner control, economic dependence and stigma. 12 Consequently, even when experiencing significant menstrual distress, they may be less likely to seek formal medical care. Whether IPV exposure modifies the fundamental relationship between menstrual symptom and healthcare-seeking remains unclear. Clarifying this relationship may help identify women with unmet needs and inform more responsive reproductive health services. In China, although IPV has received increasing attention, population-based evidence on its association with gynaecological morbidity, particularly common menstrual symptoms, remains limited. Evidence is also scarce on whether and how IPV exposure may shape healthcare-seeking pathways for PD and PMS. To address these knowledge gaps, we conducted a nationwide cross-sectional study among Chinese women. We aimed to: (1) examine the associations of overall IPV and specific IPV subtypes with PD severity and PMS and (2) investigate whether IPV exposure modifies the associations of PD severity and PMS with healthcare-seeking for menstrual problems.

We conducted a cross-sectional survey across 31 provinces in mainland China between 1 June and 20 August 2025. An online questionnaire was distributed via Wenjuanxing ( https://www.wjx.cn/ ), one of the largest online survey platforms in China. The platform maintains a nationally representative sample database comprising more than 6.2 million registered users and 10 million daily active users. Participants were recruited from this sample pool through web-based invitations, email or text messages. The full questionnaire is provided at the end of the online supplemental materials . A two-stage sampling design was used to improve national representativeness. First, quota sampling was used to allocate the initial sample size across all provincial-level administrative divisions (provinces, autonomous regions and municipalities) in mainland China, in proportion to the resident female population, according to the 2024 China Statistical Yearbook ( online supplemental table S1 ). Subsequently, within each province, potential female participants aged 18–70 years were randomly selected from the Wenjuanxing sample pool. A total of 7000 individuals were invited to participate in the survey. To further improve representativeness, post-stratification weights were applied to the final analytical sample according to the regional distribution of the female population in the 2024 China Statistical Yearbook. All data were collected through the Wenjuanxing platform. Among those invited, 6727 respondents clicked and initiated the online questionnaire. On completion, the data underwent screening against a set of pre-defined criteria (eg, age verification, logic checks and data completeness). Ultimately, 4175 participants were included in the cross-sectional analysis. Electronic informed consent was obtained before participants accessed the questionnaire. An online consent page explained the study purpose, voluntary nature of participation, anonymity, data confidentiality, potential risks and benefits and contact information for the research team. Participants could proceed to the survey only after clicking the ‘I agree to participate’ option. Those who did not provide consent were not allowed to complete the questionnaire ( figure 1 ). Exposure to IPV was assessed using a revised IPV screening tool based on the Extended-Hurt, Insult, Threaten, Scream scale. 13 – 15 The scale comprises five items evaluating the frequency of the following partner behaviours over the past year: (1) inflicting physical harm on you (eg, hitting directly or striking with an object); (2) forcing sexual activity against your will; (3) insulting or blaming you in public or private; (4) imposing controlling behaviours towards you (eg, checking your mobile phone without permission, restricting your clothing choices or social interactions) and (5) showing indifference when you were unwell. Responses for each item were scored on a 5-point Likert scale: 1=never, 2=rarely, 3=sometimes, 4=often and 5=frequently. The total score ranged from 5 to 25 and was calculated by summing the five item scores. We conducted a confirmatory factor analysis to assess the construct validity of the revised scale. The five-item IPV scale demonstrated acceptable internal consistency in this sample, with a Cronbach’s α of 0.72. Model fit was assessed using the comparative fit index (CFI), Tucker–Lewis index (TLI), root mean square error of approximation (RMSEA), and standardised root mean square residual (SRMR). The model fit indices were as follows: χ² (5)=176.45, p0.9); TLI=0.911 (>0.9); RMSEA=0.071 (<0.08), 90% CI (0.059 to 0.092); and SRMR=0.034 (<0.08). Overall model fit was acceptable. CFI, TLI and SRMR indicated satisfactory fit, while RMSEA suggested modest residual misfit but remained within an acceptable range. The standardised factor loadings for all items ranged from 0.49 to 0.69, and all were statistically significant (p<0.001), indicating that all five items effectively reflect the same underlying IPV construct. The composite reliability was 0.73, and the average variance extracted was 0.56. 14 Details are provided in online supplemental table S2 . Consistent with established criteria from prior research, a total score of ≥7 was used to define the presence of IPV exposure. 13 16 17 Furthermore, for the analysis of violence types, a score of ≥2, corresponding to ‘rarely’ or more often, on an individual item was used to define exposure to that specific type of IPV. 16 Dysmenorrhoea severity was quantified using a Numerical Rating Scale for menstrual pain, with a specific focus on PD. 18 19 To exclude the influence of underlying organic pathologies, participants were first asked whether they had a physician-diagnosed history of conditions that could cause secondary dysmenorrhoea, such as endometriosis, adenomyosis, pelvic inflammatory disease or uterine fibroids. Those confirming no such history were then instructed to recall the moment of peak pain intensity during their most recent three menstrual cycles and rate it on a continuous scale from 0 to 10, where 0 represented ‘no pain’ and 10 represented ‘the worst pain imaginable’. According to widely adopted clinical cut-offs, the pain scores were categorised as none (0), mild (1–3), moderate (4–6) and severe (7–10). This ordinal variable was used in subsequent analyses to examine the dose-response relationship between dysmenorrhoea severity and core predictors. 20 21 The presence of PMS symptoms was determined based on self-report by the study participants. With reference to established diagnostic frameworks—including the DSM-5 criteria for Premenstrual Dysphoric Disorder, 22 the International Federation of Gynecology and Obstetrics recommendations for PMS diagnosis and management, 23 the Merck Manual and domestic gynaecological clinical consensus, 24 25 we assessed 11 typical premenstrual or menstrual symptoms. Details are provided in online supplemental table S3 . These symptoms included low mood, irritability, anxiety, mood swings, headache, difficulty concentrating, breast tenderness, abdominal bloating, swelling of the extremities, hot flashes and night sweats. Together, these symptoms covered affective, cognitive and somatic domains and were considered suitable for self-report in a population-based survey. Participants indicated whether they typically experienced each symptom in the perimenstrual period. Based on this information, a binary variable indicating broadly defined self-reported PMS symptoms was created. Participants reporting at least one of the listed peri-menstrual symptoms were classified as having PMS symptoms, whereas those reporting no symptoms were classified as not having PMS symptoms. This operational definition was intended to capture a broad spectrum of premenstrual discomfort in a population-based survey, rather than to establish a clinical diagnosis of PMS. Healthcare-seeking behaviour for menstrual symptoms was assessed using the question: ‘In the past 12 months, have you sought medical care for menstrual problems, including dysmenorrhea, premenstrual syndrome and heavy menstrual bleeding?’ Responses were recorded as yes or no. Participants who reported seeking care at community health centres or higher-tiered formal medical institutions were coded as yes, and all others were coded as no. 11 26 27 Potential confounders were selected a priori based on existing literature. These included age, household registration, marital status, education level, employment status, household income, tobacco use, alcohol use, reproductive health knowledge, self-reported health and parity. Age was treated as a continuous variable. Other variables were categorised as follows: household registration, urban or rural; marital status, married or unmarried; education level, high school or below or junior college or above; employment status, unemployed, public sector or private sector; household income, low/moderate or high; tobacco use, yes or no; alcohol use, yes or no; reproductive health knowledge, low/moderate or high; self-reported health, healthy or unhealthy; and parity, yes or no. All covariates were self-reported. Descriptive statistics were used to summarise participant characteristics. Categorical variables were presented as frequencies and percentages, and continuous variables as means with SD. Initial bivariate analyses were conducted to examine crude associations between each exposure and outcome variable. Pearson’s χ² tests were used for binary outcomes, including PMS symptoms and healthcare-seeking behaviour, after confirming that cell counts were sufficient for χ² testing. For ordinal PD severity, the Mantel-Haenszel χ² test for trend was applied. Subsequently, multivariable regression models were fitted to assess adjusted associations. To account for potential clustering within the sample, all regression models were estimated using cluster-robust SEs at the provincial level. 28 Two main sets of models were developed. For the ordinal outcome of PD severity, categorised as none, mild, moderate and severe, ordinal logistic regression models were employed. These models estimated whether IPV exposure was associated with higher categories of pain severity, with results presented as ORs with 95% CIs. For the binary outcome of PMS symptoms, binary logistic regression models were used. In these models, the primary exposure was IPV status (positive/negative). All models were adjusted for the full set of covariates listed previously. 29 30 To investigate whether IPV exposure modified the relationship between menstrual symptoms and healthcare-seeking behaviour, binary logistic regression models were fitted with healthcare-seeking behaviour (binary) as the outcome, using IPV status as the grouping variable. The models included the main effects of dysmenorrhoea severity (ordinal) and PMS status (binary). Given that the decisions to seek care for dysmenorrhoea and for PMS might be correlated, seemingly unrelated regression was employed to test for coefficient differences across IPV strata, because it accounts for potential correlation between the decisions to seek care for dysmenorrhoea and for PMS, thereby providing more efficient estimates than separately testing interactions in two independent models. 31 Cluster-robust SEs were also used in this system. All regression results are presented as ORs with 95% CIs. Model assumptions were assessed using appropriate tests, including the Brant test for the proportional odds assumption in ordinal logistic regression. Where multiple comparisons were performed, adjusted p values were calculated using the Benjamini-Hochberg procedure to control the false discovery rate. All statistical tests were two-tailed, and a p value <0.05 was considered statistically significant. Analyses were performed using Stata V.18.0.

Table 1 presents the baseline characteristics of the 4175 participants included in this cross-sectional analysis. The mean age of the participants was 31.40 (SD 7.22) years. Regarding the primary exposure, 59.59% (n=2488) of participants screened positive for IPV exposure. Among IPV subtypes, psychological violence was the most commonly reported (80.10%, n=3344), followed by sexual violence (20.81%, n=869) and physical violence (10.49%, n=438). Concerning gynaecological outcomes, PD was highly prevalent, with only 15.45% (n=645) reporting no pain. The majority experienced mild pain (63.88%, n=2667), while 17.72% (n=740) and 2.95% (n=123) reported moderate and severe pain, respectively. Self-reported PMS symptoms were reported by 95.93% of participants (n=4005). This high prevalence likely reflects our inclusive operational definition, in which women reporting any of the 11 common perimenstrual symptoms were classified as having PMS symptoms. Therefore, this estimate reflects the prevalence of self-reported premenstrual discomfort rather than the prevalence of clinically diagnosed PMS. Only 23.04% (n=962) of the participants reported seeking medical care for menstrual issues at community hospitals or higher-level formal medical institutions in the past year. The sample had a relatively high socioeconomic status: 93.96% of participants (n=3923) had junior college education or above, and 75.14% (n=3137) reported a high household income. Most participants were married (71.09%, n=2968), had urban household registration (58.37%, n=2437) and were employed in the private sector (61.41%, n=2564). More than half reported unhealthy self-reported health (56.36%, n=2353) and 62.04% (n=2590) had low-to-moderate reproductive health knowledge. Data are n (%) or mean (SD). RHK, reproductive health knowledge. Online supplemental figure S1 shows the distribution and associations between key variables. IPV exposure was significantly associated with PD severity (χ²=9.20, p=0.027). Among women with no prior exposure to IPV, 16.7% reported the absence of dysmenorrhoea, compared with 14.5% in the exposed group. IPV exposure was also significantly associated with broadly defined PMS symptoms (χ²=17.72, p<0.001). PMS symptoms were reported by 96.9% of IPV-exposed women and 94.3% of non-exposed women. The severity of PD was strongly associated with healthcare-seeking behaviour (χ²=136.75, p<0.001). The proportion of women seeking medical care increased with pain severity: 14.1% among those with no pain, 21.1% with mild pain, 34.1% with moderate pain and 50.0% with severe pain. A significant association was observed between PMS and healthcare-seeking behaviour (χ²=10.05, p=0.002). Among women with PMS symptoms, 23.6% sought medical care for menstrual problems, compared with 13.2% of women without PMS symptoms. Finally, IPV exposure was significantly associated with healthcare-seeking behaviour (χ²=27.67, p<0.001), with a lower proportion of IPV-exposed women seeking medical care for menstrual symptoms than non-exposed women (19.0% vs 26.0%). To further examine whether the association between menstrual symptoms and healthcare-seeking behaviour differed by IPV exposure status, stratified analyses were performed ( online supplemental figure S2 ). The association between PMS symptoms and healthcare-seeking was significant among women not exposed to IPV (χ²=6.305, p=0.012), but not among those exposed to IPV (χ²=2.499, p=0.114). In contrast, the association between PD severity and healthcare-seeking remained significant in both strata, with a more pronounced association observed in the IPV-exposed group (χ²=81.90, p<0.001) than in the non-exposed group (χ²=52.38, p<0.001). Table 2 presents the multivariable regression results for the associations of IPV exposure with PD severity and PMS symptoms. The results indicated that IPV exposure was significantly associated with both menstrual symptoms. Women exposed to IPV had higher odds of greater PD severity (OR=1.288, 95% CI 1.133 to 1.464, p<0.001) and PMS symptoms (OR=1.979, 95% CI 1.442 to 2.715, p<0.001). Among other covariates, older age was associated with lower odds of greater PD severity (OR=0.952, p<0.001) but not significantly associated with PMS symptoms. Being married was linked to lower odds of greater PD severity (OR=0.712, p=0.006). Employment in both public and private sectors was associated with an increased risk of PD severity. Higher household income was related to a reduced risk of PD (OR=0.784, p=0.002). Tobacco use and alcohol consumption were significantly associated with higher odds of greater PD severity, while alcohol use was also associated with a higher risk of PMS symptoms (OR=1.863, p=0.018). Good self-reported health was significantly associated with a lower risk of both greater PD severity and PMS symptoms. Having a childbirth history was associated with lower odds of greater PD severity (OR=0.670, p<0.001). Model fit indices were LR χ 2 =399.59 and pseudo R 2 =0.049 for PD severity, and LR χ 2 =52.13 and pseudo R 2 =0.037 for PMS symptoms. This table presents the adjusted associations between intimate partner violence exposure, primary dysmenorrhea and self-reported premenstrual syndrome among women. The estimated models were ordinal logistic regression model and logistic regression model. *p<0.05, **p<0.01, ***p<0.001. IPV, intimate partner violence; PD, primary dysmenorrhea; PMS, premenstrual syndrome; RHK, reproductive health knowledge. In sensitivity analyses, we assessed the robustness of the findings to the operationalisation of IPV exposure by repeating the primary multivariable models with the total IPV score treated as a continuous predictor ( online supplemental tables S4 and S5 ). The results were consistent with the main analyses. Each 1-point, 3-point and 5-point increase in the total IPV score was associated with significantly higher odds of both greater PD severity (OR=1.073, 1.235 and 1.422, respectively) and PMS symptoms (OR=1.150, 1.526 and 2.022, respectively). Figure 2 and online supplemental table S6 further illustrate the associations of IPV subtypes with PD severity and PMS symptoms. After adjusting for multiple confounders, only psychological violence was significantly associated with both outcomes. Women exposed to psychological violence had 18.0% higher odds of greater PD severity (OR=1.180, 95% CI 1.005 to 1.386, p=0.043) and 78.5% higher odds of PMS symptoms (OR=1.785, 95% CI 1.258 to 2.533, p=0.001). In contrast, physical violence was not significantly associated with either PD severity (OR=1.040, p=0.720) or PMS symptoms (OR=0.715, p=0.209). Similarly, sexual violence showed no statistically significant association with PD severity (OR=1.152, p=0.089) or PMS symptoms (OR=1.508, p=0.086). Based on IPV exposure status, the sample was divided into an IPV-negative group (n=1687) and an IPV-positive group (n=2488) to explore the associations of PD severity and related factors with PMS in a stratified manner. As shown in online supplemental table S7 , PD severity was the strongest factor driving patients to seek healthcare, demonstrating a clear dose-response relationship. Compared with women reporting no or mild pain, those with moderate pain had higher odds of seeking care at formal medical institutions in both the IPV-negative group (OR=4.558, p<0.001) and the IPV-positive group (OR=3.525, p<0.001). Similar associations were observed for severe pain in the IPV-negative group (OR=6.904, p<0.001) and IPV-positive group (OR=8.692, p<0.001). Notably, a higher level of reproductive health knowledge was independently associated with greater odds of seeking healthcare for PD in both groups (p<0.001), suggesting that health awareness is a key factor promoting healthcare-seeking. As shown in online supplemental table S8 , the association between PMS symptoms and healthcare-seeking behaviour differed by IPV exposure status. PMS symptoms were significantly associated with healthcare-seeking among IPV-negative women (OR=2.357, p=0.017), but not among IPV-positive women (OR=1.672, p=0.101). Similar to the findings for PD, a higher level of reproductive health knowledge was associated with greater odds of healthcare-seeking for PMS symptoms in both groups (p<0.001). The results of the seemingly unrelated regression are presented in online supplemental figure S3 . The association between PMS symptoms and healthcare-seeking behaviour differed significantly between IPV-negative and IPV-positive women (χ²=6.61, p=0.010). In contrast, differences in the associations between PD severity and healthcare-seeking behaviour across IPV strata were not statistically significant. These findings suggest that PD severity was consistently associated with healthcare-seeking regardless of IPV exposure status, whereas the association between PMS symptoms and healthcare-seeking appeared weaker among women exposed to IPV.

To our knowledge, this is among the first population-based studies in China to examine the associations of IPV with dysmenorrhoea, PMS symptoms and healthcare-seeking for menstrual problems. Dysmenorrhoea is a health issue often overlooked by women, a point clearly illustrated by our findings. First, nearly 6 out of 10 (59.59%) participants reported exposure to IPV, a rate higher than the average reported in other studies. This discrepancy may be attributed to our expanded categorisation of psychological violence. In addition to the commonly measured dimensions of control and threats, our assessment also incorporated emotional neglect. 13 32 – 34 Psychological violence was the most frequently reported IPV subtype, with 80.10% of participants reporting exposure. This finding suggests that psychological forms of IPV may constitute a substantial component of violence exposure in this population. In terms of physical and sexual violence, the prevalence in our study is largely consistent with levels reported in previous research. 15 34 35 Second, the burden of menstrual symptoms was substantial. Broadly defined self-reported PMS symptoms were reported by 95.93% of participants, and 84.55% reported PD. Among women with PD, approximately one-fifth reported moderate or severe pain. In contrast, only about a quarter (23.04%) of women with dysmenorrhoea had sought care for such problems at community hospitals or higher-level formal medical institutions. After adjustment for sociodemographic, behavioural and reproductive health covariates, IPV exposure was correlated with elevated odds of PD and broadly defined self-reported premenstrual complaints with modest effect sizes. IPV serves as one correlate of menstrual symptom burden and healthcare-seeking patterns. Furthermore, although PD severity was a strong and consistent driver for seeking healthcare among all women, the association between PMS and healthcare-seeking was statistically significant only among women not exposed to IPV. These findings suggest that IPV exposure may be associated with a different pattern of healthcare-seeking for PMS symptoms. Additionally, a higher level of reproductive health knowledge was associated with increased odds of seeking care for both conditions, highlighting the nuanced role of health awareness. The explanatory power of the multivariable models was modest, with pseudo R 2 values below 0.05. This indicates that the included covariates explained only a limited proportion of the variation in PD severity and PMS symptoms. The low pseudo R 2 suggests that the models should be interpreted primarily as estimating adjusted associations rather than providing strong prediction or comprehensive explanation of menstrual symptom burden. Such modest explanatory power is not unexpected for menstrual symptoms, which are influenced by a complex combination of biological, psychological, behavioural, relational and healthcare-access factors. Collectively, these findings point to that a substantial number of women in China bear the dual burden of IPV exposure and significant menstrual symptoms, and their healthcare needs remain largely unmet. The findings are consistent with the WHO Global Plan of Action to End Violence against Women and Girls (2022–2030), which emphasises the need to address the health consequences of violence against women. 2 3 They may also inform future efforts related to SDG 3 on health and well-being and SDG 5 on gender equality, particularly in the areas of reproductive health and gender-based violence. 36 The high prevalence of IPV exposure, predominantly psychological violence, is consistent with recent estimates from China and global trends. 1 15 35 Psychological violence is often under-recognised yet remains highly prevalent. The observed association between IPV exposure, particularly psychological violence, with greater PD severity and PMS symptoms extends previous research that links psychosocial stress and adverse life events to worsened menstrual symptoms. 37 38 This finding underscores the profound impact of chronic emotional abuse and coercive control on women’s physical health. Psychological violence may involve sustained stress, loss of autonomy and emotional distress, which could contribute to hypothalamic-pituitary-adrenal axis dysregulation and inflammatory responses. These pathways may partly explain the observed associations with greater dysmenorrhoea severity and premenstrual symptoms. It is noteworthy that neither physical nor sexual violence showed significant associations with menstrual symptoms in our analysis. This may stem from their lower prevalence in our sample (10.49% and 20.81%, respectively) or might reflect the more pervasive and sustained impact of psychological violence on stress-mediated physiological processes. A key finding is the differing associations of IPV exposure with healthcare-seeking pathways for PD versus PMS. While PD severity remained a powerful and consistent driver for seeking care in both IPV-exposed and non-exposed groups, the motivating effect of PMS symptoms was significant only among women not exposed to IPV. This disparity aligns with WHO-emphasised ‘violence-related barriers to care’, which may be particularly influential for symptom recognition of stigmatised or ‘hidden’ conditions. PMS symptoms (eg, mood swings, anxiety) are often dismissed as ‘normal’ menstrual variations or personal failings. Within an IPV context, women’s experiences of these symptoms may be invalidated by partners or healthcare providers. 39 – 41 In contrast, the tangible and often debilitating nature of severe dysmenorrhoea may compel women to overcome such barriers to seek relief. The low rate of formal healthcare utilisation (23.04%) despite high symptom prevalence indicates a significant unmet need. This gap may be attributed to the normalisation of menstrual discomfort, lack of awareness about effective treatments, perceived ineffectiveness of healthcare services and structural barriers such as cost and time. These challenges may be further compounded in the context of IPV. Women exposed to IPV may also be reluctant to disclose PMS symptoms because these symptoms overlap with mental health concerns, disclosure could trigger further partner control or retaliation, or they may believe that healthcare providers will not address the underlying violence. 42 The low rate of formal care-seeking for menstrual symptoms (23.04%) suggests a substantial unmet health need, consistent with global estimates that only 20%–30% of women with menstrual disorders seek medical care. 43 – 45 This gap is relevant to SDG 3.7, which calls for universal access to sexual and reproductive healthcare services. 45 In China, low care-seeking may be driven by multiple factors: normalisation of menstrual pain, limited awareness of effective treatments (eg, NSAIDs for PD and non-pharmacological management for PMS), structural barriers and IPV-specific obstacles such as partner control and fear of disclosure. The consistent positive association between higher reproductive health knowledge and care-seeking across all groups underscores the potential of health education to empower women, a strategy endorsed by the WHO Global Strategy for Women’s, Children’s and Adolescents’ Health (2016–2030). 46 Our study has several limitations. First, although this nationwide cross-sectional survey provides evidence on the associations among IPV, menstrual symptoms and healthcare-seeking behaviour, its design does not allow causal inference. The temporal ordering among IPV exposure, PD severity, PMS symptoms and healthcare-seeking behaviour cannot be established. Unmeasured factors, such as prior trauma, mental health status, relationship quality, healthcare accessibility and cultural attitudes towards menstruation, may also have confounded the observed associations. Therefore, the findings should be interpreted as correlational and hypothesis-generating rather than causal. Second, all key variables were self-reported and may be subject to recall and social desirability biases, particularly for sensitive topics such as IPV. IPV exposure may have been under-reported despite the anonymous nature of the online survey, especially for physical and sexual violence. This may partly explain their lower prevalence and non-significant associations in subtype analyses. PMS was measured using a broad self-reported symptom-based definition rather than a strict clinical diagnostic framework. Although this approach helped capture a broad spectrum of perimenstrual discomfort in a population-based survey, it likely inflated the estimated prevalence of PMS symptoms and introduced heterogeneity into the PMS group. Therefore, PMS-related findings should be interpreted as referring to broadly defined self-reported premenstrual symptoms rather than clinically confirmed PMS. Third, healthcare-seeking behaviour was measured using a single binary indicator of whether participants had sought care for menstrual problems at community health centres or higher-level formal medical institutions in the past 12 months. This measure does not capture the frequency, timing, urgency, quality or perceived helpfulness of healthcare use. It also does not distinguish between different menstrual problems or between formal, informal, online, pharmacy-based and self-management responses. In addition, the multivariable models had modest explanatory power, suggesting that unmeasured biological, psychological, relational and healthcare-access factors may contribute to menstrual symptom burden and care-seeking. Future studies should use longitudinal designs, validated clinical assessments and qualitative methods to clarify temporal relationships and better understand how IPV exposure may shape menstrual health management. Despite these limitations, our findings may have implications for menstrual health services and future research. Although the observed associations were modest and the models are not suitable for individual risk prediction, the high prevalence of both IPV exposure and menstrual complaints suggests that even small differences in symptom burden or formal healthcare-seeking may correspond to substantial unmet needs at the population level. Menstrual health consultations should not be used to predict IPV or menstrual disorders, but they may provide an opportunity to identify broader psychosocial vulnerabilities and offer confidential referral and supportive services when appropriate. Menstrual health education and reproductive health literacy interventions may also help reduce the normalisation and stigmatisation of menstrual discomfort, but their effectiveness should be evaluated in future longitudinal and intervention studies.

This study identified associations between IPV exposure, menstrual symptom burden and healthcare-seeking behaviours among Chinese women. The findings indicate that social barriers to care arising from IPV, menstrual stigma and reproductive health literacy should be taken into consideration when refining menstrual health services and designing future intervention research. Longitudinal cohort studies and intervention trials are warranted to clarify the temporal sequence of relevant variables and examine whether trauma-informed reproductive health services incorporating routine IPV screening can ameliorate women’s menstrual health and improve healthcare accessibility. These efforts may contribute to broader goals of reducing gender-based violence and ensuring universal access to sexual and reproductive health services.