Family function fully mediates the relationship between social support and depression among antenatal and postpartum women in rural Southwest China

Background: Antenatal and postpartum depression is the most common complication of gestation and childbearing affecting women and their families, and good social support and family function are considered the protective and modiable factors. This study aimed to investigate the depression status and to clarify the interrelationships between social support and depression considering the inuence of family function among antenatal and postpartum women in rural areas of Southwest China. Methods: This is a cross-sectional study. Data were collected from a total of 490 rural antenatal (N=249) and postpartum(N=241) women (age: 28.17± 5.12). A structural equation modeling (SEM) was used to test the hypothesized relationships among the variables. The following instruments were used: the Edinburgh Postpartum Depression Scale (EPDS), the APGAR Family Care Index Scale, and the Social Support Rate Scale (SSRS). Results: We found that the prevalence of depression was 10.4%. Path analysis showed that family function had a direct negative correlation with depression (β =-0.251, 95%CI: -0.382 to -0.118). Social support had a direct positive correlation with family function (β =0.293, 95%CI: 0.147 to 0.434) and had an indirect negative correlation with depression (β =-0.074, 95%CI: -0.139 to -0.032), family function fully mediated the relationship between social support and depression. Conclusions: Findings of this study highlight that family function should be considered as the key target for interventions aiming to lower the prevalence of antenatal and postpartum depression. Family members interventions are critical to reduce depression among antenatal and postpartum women.


Background
Perinatal period is special for women as a major family transition, easily leading to the onset and recurrence of mental illness, and depression and anxiety are the most common [1,2]. Perinatal depression is of concern not only because of the suffering and distress it cause for women, also because of the risk of adverse effects on the developing fetus and child [3].The prevalence of perinatal depression in China is 16.3% reported in a meta analysis conducted by Anum Nisar in 2020, and an increasing trend was seen over the last decade and in underdeveloped regions [4].
The prevalence of antenatal depression ranges from 5.2% to 17.8% worldwide [5], which can be partly explained by the negative effects of uctuations in gonadotropin levels during pregnancy on neurotransmitter levels and functional differences in the hypothalamus-pituitary-adrenal axis [6]. Women with antenatal depression appear to be at considerably higher risk for self-harm or suicide, failure to seek prenatal care, poor diet [7,8], which may lead to the adverse pregnancy outcomes(like complications during pregnancy, premature birth, dysplasia of the fetus, low birth weight of the baby) [9,10]. Moreover, a study had shown that women were more vulnerable to psychiatric illness after birth [11]. The prevalence of postpartum depression is 14.8% in China and there is a rising trend [4]. As well, postpartum depression also predicts shorter breastfeeding time [12] and may be a risk factor for children with low social capacity [13]. Existing literature shows that factors associated with pregnancy and postpartum depression mainly include sociodemographic characteristics(e.g. age, marital status, education, income, complications of pregnancy) [14][15][16][17], social support [18], and family function [19].
Social support is de ned as instrumental, informational and emotional support provided by a social network including family, friends and neighborhoods, which could protect psychological well-being through buffering the effects of traumatic life events [20]. It can be characterized by the provider of support, including support from a spouse, relatives or friends, each thought to have independent protective effects against depression [21]. Social support as a protective and modi able factor has been well investigated in relation to antenatal or postpartum depression [22]. Some studies have also proved that low-level social support were risk factors for antenatal and postpartum depression [23,24]. And then, family ties will strengthen, same as the Tarkka's and Lepistö S's studies that social support has been considered an important resource to improve family functioning [25,26].
Family function can be de ned as the degree to which a family performs as a unit to manage conditions, activities, external stimuli or events that cause stress [27]. Compared to healthy families, families with family dysfunction are expressed as lower cohesion, lower warmth, and expressiveness and higher con ict, rigidity, and affectionless control [19]. Previous studies had shown that depression is negatively correlated with family functioning [28,29]. Improvement of family function may contribute to better prognosis of depression [30]. In addition, Wang Y et al (2019) proposed a model that family function moderated the indirect relationship between social support and depression in the elderly [31].
The availability of mental health resources in rural areas of China is low [32]. Studies have shown that living in rural areas of China strongly signi cantly associated with antenatal and postpartum depression [33]. Despite previous studies have demonstrated the relationship between family function and depression as well as social support and depression, few studies have included these three variables in one study to understand the interrelationships and potential mechanisms of social support, family function and depression among antenatal and postpartum women. We examined the in uence of social support and family function on the antenatal and postpartum depression in rural areas of southwest China in this study. Based on the above description, we hypothesized a single mediator model shown in Figure 1. Speci cally, social support would be positively associated with family function (hypothesis 1) and negatively associated with depression (hypothesis 2). We also hypothesized that the family function would be negatively associated with depression (hypothesis 3). In addition, we suggested that the relationship between social support and depression would be mediated by the family function (hypothesis 4). The study objects to assess the prevalence of antenatal and postpartum depression in rural China and identify the key factors including social support and family function which contribute to the prevention and control of depression in antenatal and postpartum women.

Participants and Procedure
This cross-sectional study was conducted among pregnant and postpartum women in rural areas of Sichuan Province, Southwest China, from December, 2017 to May, 2018. The optimal time to conduct the rst screen for postpartum depression is within 6 months postpartum [34,35]. So, the target population in this study was the women who were at pregnancy or within 6 months postpartum.
A multi-stage strati ed random sampling was used to acquire the sample. In the rst stage, we randomly chose a city in Sichuan province. In the second stage, we randomly selected a rural district in the city. In the third stage, 10 townships were randomly selected from the rural district. In the fourth stage, we randomly selected 50 maternal women from the database of maternal women established by each township hospital. Trained investigators invited the selected participants to take part a face to face interview in their home and the questionnaires were completed by the investigators. We used the quanti able scales, trained investigators, two-person data entry and logical veri cation to ensure the quality of the research.

Ethical Consideration
The study protocol was approved by the Institutional Review Board of Sichuan University (Project identi cation code: H171260). The study was explained to participants and informed written consent was obtained prior to data collection.

Measures
Participants' socio-demographic characteristics, social support, family function and depression information were collected from questionnaires.

Socio-demographic Characteristics
Socio-demographic Characteristics included age, type of maternal women, marital status, education level, employment status, individual annual income, medical insurance status and complications of pregnancy.

Social support
Social support was assessed through the Social Support Rating Scale, which was developed by Xiao S.Y et al [36]. The SSRS was speci cally designed for use in a Chinese context and consists of ten items of three domains in total: objective support, subjective support and social support utilization. Responses were provided as a 4-point Likert scale, the overall score of all items ranges from 12 to 66 with higher scores re ecting stronger social support. The total score has been divided into three levels: low (12)(13)(14)(15)(16)(17)(18)(19)(20)(21)(22), moderate , and high . The SSRS has been widely applied in China with excellent validity and reliability [37,38].In this research, Cronbach's α of the scale was 0.825.

Family function
Family function was measured by the APGAR, developed by Smilkstein, et al. [39], which was used to evaluate an individual's satisfaction with family function. This scale was a 3-point scale ranging from 0 (hardly ever) to 2 (almost always), composing of ve items: adaptation, partnership, growth, affection and resolve. The total score ranged from 0 to 10 with higher scores denoting higher level of satisfaction with family function. It was generally believed that score 0-3 indicated severe family dysfunction, 4-6 indicated moderate family dysfunction, and 7-10 indicated good family function. The Chinese version of APGAR has been widely applied in China with excellent validity and reliability [40,41]. In this research, Cronbach's α of the scale was 0.874.

Depression
Depression was measured by EPDS (Edinburgh Postnatal Depression Scale). The EPDS, designed by Cox, et al. [42], was originally developed to assist primary care health professionals to detect mothers suffering from postpartum depression and was also proved to be suitable for the detection of antenatal depression in 2003 [43].The EPDS is a 10-item self-reported questionnaire on depressive symptoms. Each item is scored on a 4-point scale (from 0 to 3), so that the total score ranges from 0 to 30, with higher scores representing more depressive symptoms. The EPDS was translated into Chinese version by Pen et al in 1994 [44], who recommended that the cut-off score for the Chinese was 9.5, and the score of 9.5 or higher indicates signi cant depressive symptoms. In this research, Cronbach's α of the scale was 0.776.

Statistical analyses
The data were entered using the Epidata3.1 database and were analyzed using the SPSS version 20.0 (SPSS Inc., Chicago, IL, USA) and Analysis of Moment Structures (AMOS) version 24.0 (IBM, New York, NY, USA). First, we calculated descriptive statistics (frequencies, percentages, means, and standard deviations) to examine the socio-demographic characteristics of the sample. Second, we undertook a descriptive analysis of study variables (means and standard deviations). Third, binary logistic regression models were used to test the relationship between social support, family function and depression. In model 1, we used depression as the dependent variable and social support, socio-demographic variables as independent variables. In model 2, we further added family function as an independent variable.
Fourth, a structural equation model (SEM) was employed to further test the hypothesized relationships among social support, family function and depression of antenatal and postpartum women.
The SEM used bootstrap maximum likelihood estimation and the results, with a p-value of < 0.05, were considered statistically signi cant. To examine the model t, we employed several indicators with their cutoffs: adjusted goodness of t index (AGFI), a goodness of t index(GFI), the comparative t index (CFI), normed t index (NFI), incremental (IFI), and Tucker-Lewis index (TLI) of 0.90 or above; a root mean squared error of approximation (RMSEA) less than or equal to 0.08, indicated an acceptable model t [45].

Participants and Socio-demographic characteristics
498 participants out of a total of 500 participants in these townships returned questionnaires (a response rate of 99.6%). Questionnaires were checked after the interviews for completeness. Eight records met exclusion criteria (Incomplete data collection: n=2; postpartum period>6 months n=6). Overall, 490 questionnaires were valid.
Socio-demographic characteristics of the 490 samples are shown in Table 1. The proportion of antenatal women and postnatal women were 50.8% and 49.2%, respectively. The mean age was 28.17±5.12, ranged from 19 to 43 years old. Most were married (96.7%), educated high or vocational school or less (73.7%). The majority of the women were currently unemployed (57.3%), had an individual annual income less than $750 (41.2%) and received medical insurance (98.0%). Most had no complications of pregnancy (81.2%). Table 2 shows scores of social support, family function and depression. The mean score of social support was 40.79 ± 5.95,0.2% (1), 71.6% (351) and 28.2% (138) of participants had low, moderate and high social support, respectively. The mean score of family function was 8.80 ± 1.89, 85.5% (419) of participants had good family function, 13.1% (64) and 1.4% (7) Table 2 shows scores of social support, family function and depression. The mean score of social support was 40.79 ± 5.95,0.2% (1), 71.6% (351) and 28.2% (138) of participants had low, moderate and high social support, respectively. The mean score of family function was 8.80 ± 1.89, 85.5% (419) of participants had good family function, 13.1% (64) and 1.4% (7)   The estimates for direct, indirect and total effects with bias-corrected 95% CI are shown in Table 4. In these analyses, effect coe cients were substantially signi cant if the 95% CI does not include 0. The results showed that social support had a signi cant positive correlation with family function (β =0.293, 95%CI: 0.147 to 0.434), thus supporting Hypothesis 1. However, the direct impact of social support on depression proved to be statistically non-signi cant (β =-0.090, 95%CI: -0.213 to 0.043), leading us to reject Hypotheses 2. Family function had a direct negative correlation with depression (β =-0.251, 95%CI: -0.382 to -0.118), thus supporting Hypothesis 3. In addition, social support had an indirect negative correlation with depression (β =-0.074, 95%CI: -0.139 to -0.032), thus supporting Hypothesis 4.

Descriptive analysis of study variable
Regarding the path between social support and depression, total effect and indirect effect are statistically signi cant but the direct effect is statistically non-signi cant. Based on the above, family function fully mediates the relationship between social support and depression.

Discussion
This study aimed to investigate the depression status and to clarify the interrelationships between social support and depression considering the in uence of family function among antenatal and postpartum women in rural areas of Southwest China. To the best of our knowledge, this is the rst study reporting the fully mediating role of family function between social support and depression in antenatal and postpartum women. Findings of this study have important implications for the development and implementation of intervention strategies and measures to ameliorate maternal depression as well as the promotion of mother's health and the future wellbeing of their children and family in rural areas.
The antenatal depression prevalence in our study is 10.4%, which is close to a meta-analysis of 21 studies(10.7%) [46]. And the prevalence rate of postpartum depression in our study is also 10.4%, lower than that of a prospective cohort study in China(19.9%) conducted by Xie R.-H et.al [38]. Different sample characteristics may be the reason. Postpartum women in our study were within 6 months, while Xie R.-H's were within one month. Women during the rst month postpartum are required to engage in speci c practices to promote the health of the maternal/newborn dyad for the Chinese tradition of "doing the month" which mainly including the promotion of maternal rest, discouraging domestic duties and activity outside the home [47]. Due to the physical and social activity limitation, combined with the frustration for breastfeeding and the lack of sleep cannot be relieved effectively, the women would be stressful which lead to mood alterations [48]. After the rst month, with social activities resumed, they may accommodate and accept to the new situation gradually [49].
The mean family function (APGAR) score was 8.80±1.89, and only 1.4% women reported severe family dysfunction in our study. The possible reason may be that in traditional rural China, women 's pregnancy is regarded as a great familial contribution, and then family members will acknowledge the women's family status and strive to develop better family functions [50].The model revealed that antenatal and postpartum women with lower family function were more likely to experience depression symptoms, which is consistent with previous studies [29,51]. There are two possible explanations. One possible explanation is that the couple relationship which plays an important role in family function, was affected after the birth of a child by the increased con ict [52] and less opportunities for shared intimacy, thus leading to the women's worse mood [53]. Another possible explanation is that women in dysfunctional families are less able to communicate their emotions and thoughts effectively with other family members, thus leading to the development of depression [54].
The results revealed that the mean score of social support among antenatal and postpartum women was 40.79±5.95, which is lower than another study(43.34±7.06) in China that surveyed women before pregnancy [55].The possible reason may be that the women after pregnancy would decrease physical exercise and leisure activities due to the concerns of maternal/child health, and thus receive less social support [56]. Our model revealed that better social support predicted better family function which is consistent with Jiang H's study [57]. There are two possible reasons. One reason may be that social support can improve physical health by increasing healthy activities, protective behaviors and promoting a healthier lifestyle, the individual and family functions get better accordingly [58]. Another reason may be that in China the family members normally provide the most solid support in one's social network, good social support usually means good family function [59].
The most signi cant nding of this study was that the relationship between social support and depression was fully mediated by family function. Previous studies identi ed that the social support had direct effect on depression [23,60], but this research further found the effect was indirect. Our model reveals that the higher social support among antenatal and postpartum women were less likely to experience depression symptoms which is consistent with previous studies, but interestingly, the association was fully mediated by family function. Family function is the key factor. This can be explained by the vulnerability-stress model, when antenatal or postpartum women facing the stressor, the low social support leads to family dysfunction which increases environmental vulnerability and triggers the onset of depression [61]. Compared to western women, Chinese women seem to be more familyoriented and thus are more likely to be affected by family relationships [62].
Findings of this study highlight the importance of family function in decreasing depression of rural antenatal and postpartum women and have important implications for public health practices. Healthcare professionals should pay more attention to evaluate family function constantly across the perinatal period and take partner-inclusive intervention to lower risk of antenatal and postpartum depression [63]. Combining assessments like APGAR, especially applying the simple Resolve item-"Are you satis ed with the way you and your family share time together?", can help professionals quickly assess family function [64]. For the dysfunctional family, the health professionals should focus on intervening in family members in addition to antenatal and postpartum women, such as requiring family members to participate in prenatal health and baby care education, providing different types of health education programs for different family members and setting up consulting platforms of perinatal nursing for families [65]. They are bene cial for minimizing the harmful effects of family dysfunction.
Limitations of this study need to be recognized. Firstly, we cannot make claims about causality among the three variables because of the cross-sectional design. Future longitudinal or experiment studies should be conducted to provide causal inference. Secondly, some factors such as life stress, personal history of depression and family history of depression have not been taken into consideration, which may also in uence the depression level among antenatal and postpartum women. In addition, the EPDS is a screening tool rather than a diagnostic tool, which can only provide information on symptoms of depression. Finally, although our study concerned with people in the community, which could reduce selection bias, the data were obtained in rural areas of southwest China, so we should be careful to generalize the ndings.

Conclusions
The study investigates the interplay between social support and depression considering the in uence of family function. The results suggested that family function played a fully mediating role in the association between social support and depression. Findings of this study highlight that family function should be considered as the key target for interventions aiming to lower the prevalence of antenatal and postpartum depression. Family members interventions are critical to reduce depression among antenatal and postpartum women.

Declarations
Ethics approval and consent to participate: The study protocol was approved by the Institutional Review Board of Sichuan University (Project identi cation code: H171260). The study was explained to participants and informed written consent was obtained prior to data collection.

Consent to publication: Not applicable
Availability of data and materials: The datasets used and/or analyzed during the current study available from the corresponding author on reasonable request.  Notes: AOR means adjusted odds ratio, ** p < 0.05. Table 4. Direct, indirect and total effects and 95% confidence intervals for the final model.