Association between social support and postpartum depression

Postpartum depression is common; however, little is known about its relationship to social support and postpartum depression. This study examined the association between them among South Korean women within one year of childbirth. This study was based on the 2016 Korean Study of Women’s Health-Related Issues (K-Stori), a cross-sectional survey employing nationally-representative random sampling. Participants were 1,654 postpartum women within a year of giving birth. Chi-square test and logistic regression analysis were conducted to analyze the associations between social support (and other covariates) and postpartum depression. Among participants, 266 (16.1%) had postpartum depression. Depending on the level of social support, 6.0%, 53.9%, and 40.1% of them had low, moderate, and high social support, respectively. Women with moderate or low social support were more likely to have postpartum depression (OR = 1.78, 95% CI = 1.26–2.53; OR = 2.76, 95% CI = 1.56–4.89). This trend was observed in participants with multiparity, pregnancy loss, obese body image, and employed women. Social support was associated with a decreased likelihood of postpartum depression, indicating the importance of social support, especially for women experiencing multiparity, pregnancy loss, negative body image, as well as for employed women.

www.nature.com/scientificreports/ perceived social support, which subjectively measures an individual's beliefs about the available support. It reflects an individuals' feeling that they are accepted, loved, and valued by other members of their social network 34,35 . Previous studies have investigated the relationship between PPD and social support 2,10,12 However, there was relatively little consideration for factors affecting social support and postpartum depression. Therefore, the aim of this study was to estimate the prevalence of PPD in postpartum women in South Korea, and investigate the relationship between PPD and social support.

Results
A total of 1,654 postpartum women were included in the analysis, of which 266 (16.1%) had PPD and 1,388 (83.9%) did not. The demographic and health characteristics of the study participants are summarised in Table 1. Statistically significant differences were observed between social support and having PPD. There were a greater number of women with PPD in groups with low social support. Table 2 shows the social support status among participants: 6.0% of women had low, 53.9% had moderate and 40.1% had high social support. Overall, women who responded that social support was low had a low level of education and income, were employed, and faced high parenting burden during the previous month (p < 0.001). In addition, women with lower level of social support had relatively higher level of stress, with a history of depression and smoking (p < 0.001). Meanwhile, women with high social support had high breastfeeding percentages and high level of subjective health status (p < 0.001). Univariate and multiple logistic regression were conducted to determine the association between PPD and social support while controlling for potential covariates (Table 3). In univariate logistic regression analyses, the women with moderate (OR = 2.66, 95% CI = 1.93-3.67) and low (OR = 6.89, 95% CI = 4.22-11.24) level of social supports had increased likelihood for PPD compared to the women with high level of social support. In multivariate logistic regression analyses, the women with moderate and low social support levels were 1.78 (95% CI = 1.25-2.52) and 2.73 (95% CI = 1.54-4.83) times more likely to develop PPD, respectively, as compared to the women with high social support levels. Further, women who were employed, breastfed, faced high parenting burdens during the previous month, had poor subjective health status, high stress level, and a past depression diagnosis were more likely to develop PPD.
The results of subgroup analyses on social support and PPD with covariates are summarised in Table 4. Among women with multiparity, those with moderate and low levels of social support were 2.85 (95% CI = 1.68-4.82) and 4.90 (95% CI = 2.14-11.23) times more likely to develop PPD, respectively; but there were no statistical differences in women with primiparity. Further, in women who experienced pregnancy loss, those with lower levels of social support were 10.26 times more likely to develop PPD compared to women receiving high levels of social support. Among women who reported their body image as normal or obese, women with moderate and low levels of social support were more likely to develop PPD. In addition, women with jobs and low levels of social support showed the highest likelihood of PPD (OR = 10.34, 95% CI = 2.34-45.64). Figure 1 presents the relationship between the level of social support in each scale of the development of PPD, and shows that the lower the social support level in all subscales, the higher the odds ratios (ORs) of PPD. The results showed a high degree of association in order of family, significant others, and friends.

Discussion
The World Health Organization (WHO) noted that mental health problems such as depression and anxiety are common during pregnancy and after childbirth 26 . However, there was limited information on which women develop PPD. Therefore, this study was conducted to investigate the associated factors of PPD, especially how social support affects depression in postpartum women. Our findings show that postpartum women with low social support had 4.63-fold higher odds of PPD compared with postpartum women with high social support. Further, women who are employed or breastfeeding, or have heavy parenting burden, poor subjective health status, high stress level, and were diagnosed with depression in the past were more likely to develop PPD. Furthermore, this study also showed that higher levels of social support may buffer against probability of PPD after adjusting for confounding variables. Given our results, postpartum women need a high level of social support from family, close friends, and significant others.
Interestingly, the subgroup analysis showed that women with multiparity had a five times higher risk of PPD if their social support was low. Multiparity could increase the level of maternal stress and depression because women need to also care for their previous children and infants. Women with multiparity may not receive the same level of social support as they received during their first time of childbearing because they are considered to be child-care experts despite their need of extra social support to take care of their new-born babies. Thus, the results suggest that as women with multiparity are more likely to have PPD, better social support is needed for preventing PPD 27 .
Another interesting finding in the current study is that women with low social support and previous experience of pregnancy loss were 10 times more likely to develop PPD. Our result is in line with previous studies in which previous pregnancy loss served as an effect modifier between social support and PPD. Pregnancy loss is an event which makes bereaved women particularly prone to depression, mood disorders, dramatic mental health disorder, and even suicide 36 . According to previous research, women who have lost their babies were seven to nine times more vulnerable to depression than women without a pregnancy loss 37 . These results indicate that postpartum women with history of pregnancy loss need higher level of social support to prevent PPD.
Body image also acts as an effect modifier between social support and PPD. Similar to adolescence, the period surrounding childbirth is accompanied by unique and rapid changes in not only body shape and size, but also psychological dimensions. Prior studies suggested that about 85% of women during pregnancy experience body image dissatisfaction 38 . In recent years, an increasing number of women reported to be concerned about their weight gain and appearance during pregnancy and postpartum period 39 www.nature.com/scientificreports/ and health behaviours in postpartum women 41 . The current study results suggest that social support is an important factor that improves mental health of women with negative body image during postpartum period. Several countries have implemented paid leave to help working parents 42 . Despite policies like paid leaves, women continue working for reasons such as career, being worried about losing their jobs, financial burden, and negative attitudes in the workplace. Women who have to multitask by having a job and having to rear child are especially vulnerable PPD 43 . Similarly, the current study results showed that employed women were 2.8 times more likely to develop PPD. Job strain often cannot be decreased or relieved, but social support in workplace can buffer the negative effect of overwork and role ambiguity 25,43 . In addition to workplace support, social support from their partner, family, and friends also decrease job stress in postpartum women 44 .
The strength of the study is that the findings are based on a nationwide survey. This ensures that the data is representative of Korean women, and comprehensively assesses depressive symptoms of the postpartum period.

Limitations
Several limitations to the present study could influence the interpretation of our findings. First, the cross-sectional design of K-stori could not show the direction of the causal relationship for the identified association between social support and PPD. Second, this study primarily relied on self-report measures from K-stori. Thus, questions on the dependent and other independent variables may contribute to recall bias. Third, depression was measured based on a self-report. Previous research has shown that self-reported survey can be under or overestimated depending on individual characteristics. Though all the responses were anonymous, quite a few individuals may not indicate their true levels of depression 45 . This study provided a cross-sectional estimate of PPD within one year of childbirth in South Korean women. Social support was also positively associated with a lower likelihood of PPD in women with multiparity, pregnancy loss experience, negative body image, and jobs. Postpartum women should receive a high level of social support from family, friends, and significant others to prevent PPD and improve their maternal health, aided by health professionals.  46 . An approved study description was provided to all eligible participants. The study description covered the research purpose, subject, content, duration, voluntary participation, withdrawal of consent, expected risks and benefits from participating in the research, publication of the study results, and confidentiality. If the subjects agreed to participate in the study after reading the study description, participants were asked to provide written informed consent 46 . All methods were carried out in accordance with approved guidelines and regulations. The participants of this study were women within a year of giving birth. From the total pregnant and postpartum women (n = 3000), pregnant women (n = 1,346) were excluded, and thus 1,654 women aged 19-44 years were included in the study. To recruit the participants, the interviewers planned to visit obstetrics and gynecology, www.nature.com/scientificreports/ or post-partum care centers. In order to select the same survey area as the other life cycle stages, a systematic data extraction method was used to identify local obstetrician and post-partum care centers, for pregnant and post-partum women, based on the same sample design area for the household survey participants.
Measures. The main outcome variable of this study was PPD, which was evaluated using the Edinburgh Postnatal Depression Scale (EPDS) 47 . The EPDS is a validated 10 questions screening tool, and is the most widely used screening questionnaire asking mothers how they have felt in the past seven days for PPD (Appendix 1). Participants answered the questionnaires and the answers were scored from 0-3 points (or 3-0 in case of a reverse score) with a total score of 0 to 30. A threshold score of ≥ 10 was used to classify postpartum women with a probable major depression who needed further medical examination based on the Korean version of the EPDS 13,48,49 . The Multidimensional Scale of Perceived Social Support (MSPSS) was used to measure individual perceived social supports from three sources: friends (Item 6, 7, 9, and 12), family (Item 3, 4, 8, and 11) and significant www.nature.com/scientificreports/ others (Item 1, 2, 5, and 10) 50 . Participants were asked to indicate their agreement with the statements on a fivepoint Likert scale ranging from 0 = very strongly disagree to 4 = very strongly agree. The total score was calculated as the mean of 12 scores. The subscale total scores were the sum of the scores for the four questions related to the subscale. Total scores ranging from 12 to 24 were classified as low social support, from 25 to 36 as moderate social support and from 37 to 48 as high social support. Further, information on age, living area, education level, household income, current job, number of parities, pregnancy loss experience, current breastfeeding status, degree of parenting burden within the last month, subjective health status, perceived body image, stress, past diagnosis of depression, and smoking experience was collected.
Statistical analyses. Descriptive analysis was conducted to compare the characteristics of study participants according to postpartum depression level. Multiple logistic regression was used to determine the association between PPD and social support while controlling for potential covariates. The OR and 95% confidence interval (95% CI) of having PPD were estimated. P values < 0.05 were considered statistically significant. Furthermore, subgroup analyses were conducted to assess the influence of social support on the risk of having PPD according to groups with different characteristics. In the subgroup analyses, the Cochran-Armitage test was used to assess the association between PPD and each variable, and the awareness of nutrition labelling. P values < 0.05 were considered to be statistically significant. All statistical analyses were performed using SAS version 9.4 (Cary, NC, USA).