Advertisement

Postpartum Perceived Stress Explains the Association between Perceived Social Support and Depressive Symptoms

  • Krista S. Leonard
    Affiliations
    Exercise Psychology Laboratory, Department of Kinesiology, The Pennsylvania State University, University Park, Pennsylvania
    Search for articles by this author
  • M. Blair Evans
    Affiliations
    Department of Kinesiology, The Pennsylvania State University, University Park, Pennsylvania
    Search for articles by this author
  • Kristen H. Kjerulff
    Affiliations
    Department of Public Health Sciences, Penn State College of Medicine, Hershey, Pennsylvania

    Department of Obstetrics and Gynecology, Penn State College of Medicine, Hershey, Pennsylvania
    Search for articles by this author
  • Danielle Symons Downs
    Correspondence
    Correspondence to: Danielle Symons Downs, PhD, Professor of Kinesiology and Obstetrics and Gynecology, Department of Kinesiology, The Pennsylvania State University, 266 Recreation Building, University Park, PA 16802-5701. Phone: (814) 863-0456; fax: 814-865-1275.
    Affiliations
    Exercise Psychology Laboratory, Department of Kinesiology, The Pennsylvania State University, University Park, Pennsylvania

    Department of Obstetrics and Gynecology, Penn State College of Medicine, Hershey, Pennsylvania
    Search for articles by this author

      Abstract

      Background

      Limited research has focused on longitudinal interrelations between perceived social support, perceived stress, and depressive symptoms beyond the first postpartum months. This study tested an alternative primary hypothesis within the stress process model examining whether perceived stress mediated the association between perceived social support and depressive symptoms from 1 to 24 months postpartum. Secondary purposes examined whether these factors 1) changed from 1 to 24 months postpartum and 2) predicted depressive symptoms.

      Methods

      Women (N = 1,316) in a longitudinal cohort study completed validated measures of perceived social support, perceived stress, and depressive symptoms at 1, 6, 12, 18, and 24 months postpartum via telephone interviews. Analyses examined changes in psychosocial factors (repeated measures analysis of variance) and the extent to which perceived social support and perceived stress predicted depressive symptoms and supported mediation (linear regression).

      Results

      Perceived social support decreased, perceived stress increased, and depressive symptoms remained constant from 1 to 18 months, then increased at 24 months. Low perceived social support predicted 6-month depressive symptoms, whereas perceived stress predicted depressive symptoms at all time points. Perceived stress mediated the association between perceived social support and depressive symptoms across 24 months such that low perceived social support predicted perceived stress, which in turn predicted depressive symptoms.

      Conclusions

      Intervention scientists may want to focus on strengthening perceived social support as a means to manage perceived stress in an effort to prevent a long-term trajectory of depression.
      Postpartum depression, a major depressive episode within the first four weeks after childbirth, is experienced by 11% to 17% of mothers world-wide (
      American Psychiatric Association
      Diagnostic and Statistical Manual of Mental Disorders.
      ,
      • Shorey S.
      • Chee C.Y.I.
      • Ng E.D.
      • Chan Y.H.
      • Tam W.W.S.
      • Chong Y.S.
      Prevalence and incidence of postpartum depression among healthy mothers: A systematic review and meta-analysis.
      ,
      • Woody C.A.
      • Ferrari A.J.
      • Siskind D.J.
      • Whiteford
      • Harris M.G.
      A systematic review and meta-regression of the prevalence and incidence of perinatal depression.
      ). Actual prevalence, however, may be higher given estimates show that more than half of women with postpartum depression often go undiagnosed (
      • Thurgood S.
      • Avery D.M.
      • Williamson L.
      Postpartum depression (PPD).
      ). Postpartum depression has detrimental effects on the mother (e.g., chronic depression;
      • McCall-Hosenfeld J.S.
      • Phiri K.
      • Schaefer E.
      • Zhu J.
      • Kjerulff K.
      Trajectories of depressive symptoms throughout the peri- and postpartum period: Results from the First Baby Study.
      ) and the infant (e.g., breastfeeding problems, maternal–infant bonding) as well as mother–partner bonding (
      • Becker M.
      • Weinberger T.
      • Chandy A.
      • Schmukler S.
      Depression during pregnancy and postpartum.
      ,
      • Behrendt H.F.
      • Konrad K.
      • Goecke T.W.
      • Fakhrabadi R.
      • Herpetz-Dahlmann B.
      • Firk C.
      Postnatal mother-to-infant attachment in subclinically depressed mothers: Dyads at risk?.
      ,
      • Dias C.C.
      • Figueiredo B.
      Breastfeeding and depression: A systematic review of the literature.
      ,
      • Mamun A.A.
      • Clavarino A.M.
      • Najman J.M.
      • Williams G.M.
      • O’Callaghan M.J.
      • Bor W.
      Maternal depression and the quality of marital relationship: A 14-year prospective study.
      ). Thus, research and public health efforts are needed to identify modifiable factors to decrease depressive symptoms and develop effective and timely interventions to manage and/or prevent the onset of symptoms.
      Many women have reported feeling lonely after birth because the general focus is on the newborn baby and not the mother (
      • American College of Obstetricians and Gynecologists
      Optimizing postpartum care.
      ,
      • Christenson A.
      • Johansson E.
      • Reynisdottir S.
      • Torgerson J.
      • Hemmingsson E.
      Women’s perceived reasons for their excessive postpartum weight retention: A qualitative interview study.
      ). Subsequently, some women report experiencing low levels of perceived social support and high levels of perceived stress during the postpartum period (
      • Li Y.
      • Long Z.
      • Cao D.
      • Cao F.
      Social support and depression across the perinatal period: A longitudinal study.
      ,
      • Robertson E.
      • Grace S.
      • Wallington T.
      • Stewart D.E.
      Antenatal risk factors for postpartum depression: A synthesis of recent literature.
      ). Low perceived social support is associated with postpartum depressive symptoms and this association can be apparent up to 24 months postpartum (
      • Ghaedrahmati M.
      • Kazemi A.
      • Kheirabadi G.
      • Ebrahimi A.
      • Bahrami M.
      Postpartum depression risk factors: A narrative review.
      ,
      • Li Y.
      • Long Z.
      • Cao D.
      • Cao F.
      Social support and depression across the perinatal period: A longitudinal study.
      ,
      • Milgrom J.
      • Hirshler Y.
      • Reece J.
      • Holt C.
      • Gemmill A.W.
      Social support – A protective factor for depressed perinatal women.
      ,
      • Norhayati M.N.
      • Nik Hazlina N.H.
      • Asrenee A.R.
      • Wan Emilin W.M.A.
      Magnitude and risk factors for postpartum symptoms: A literature review.
      ,
      • Reid K.M.
      • Taylor M.G.
      Social support, stress, and maternal postpartum depression: A comparison of supportive relationships.
      ,
      • Schwab-Reese L.M.
      • Schafer E.J.
      • Ashida S.
      Associations of social support and stress with postpartum maternal mental health symptoms: Main effects, moderation, and mediation.
      ,
      • Tambag H.
      • Turan Z.
      • Tolun S.
      • Can R.
      Perceived social support and depression levels of women in the postpartum period in Hatay, Turkey.
      ). Many women express the desire for more support to help with adjusting to new roles as a mother and caretaker as well as managing the physical and psychological changes to their bodies after delivery (
      • Christenson A.
      • Johansson E.
      • Reynisdottir S.
      • Torgerson J.
      • Hemmingsson E.
      Women’s perceived reasons for their excessive postpartum weight retention: A qualitative interview study.
      ). Women have reported that when these needs are not met, it increases experiences of postpartum depressive symptoms (
      • Christenson A.
      • Johansson E.
      • Reynisdottir S.
      • Torgerson J.
      • Hemmingsson E.
      Women’s perceived reasons for their excessive postpartum weight retention: A qualitative interview study.
      ). Moreover, the postpartum period can introduce women to daily life stressors not previously experienced (e.g., caretaking for a newborn, sleep disturbances, balancing work and life responsibilities), increasing overall levels of perceived stress. This high level of perceived stress during the postpartum period may elevate or augment depressive symptoms (
      • Ghaedrahmati M.
      • Kazemi A.
      • Kheirabadi G.
      • Ebrahimi A.
      • Bahrami M.
      Postpartum depression risk factors: A narrative review.
      ,
      • Grote N.D.
      • Bledsoe S.E.
      Predicting postpartum depressive symptoms in new mothers: The role of optimism and stress frequency during pregnancy.
      ,
      • Norhayati M.N.
      • Nik Hazlina N.H.
      • Asrenee A.R.
      • Wan Emilin W.M.A.
      Magnitude and risk factors for postpartum symptoms: A literature review.
      ,
      • Razurel C.
      • Kaiser B.
      • Antonietti J.P.
      • Epiney M.
      • Sellenet C.
      Relationship between perceived perinatal stress and depressive symptoms, anxiety, and parental self-efficacy in primiparous mothers and the role of social support.
      ,
      • Robertson E.
      • Grace S.
      • Wallington T.
      • Stewart D.E.
      Antenatal risk factors for postpartum depression: A synthesis of recent literature.
      ,
      • Schwab-Reese L.M.
      • Schafer E.J.
      • Ashida S.
      Associations of social support and stress with postpartum maternal mental health symptoms: Main effects, moderation, and mediation.
      ). Given the increased risk of postpartum depressive symptoms owing to poor perceived social support and high perceived stress, researchers have attempted to better understand the factors as a whole and examine their complex interrelations.
      To this end, past studies have examined perceived social support, perceived stress, and postpartum depressive symptoms under the stress process model, whereby perceived social support is a mediator of the association between perceived stress and depressive symptoms (
      • Pearlin L.I.
      • Lieberman M.A.
      • Menaghan E.G.
      • Mullan J.T.
      The stress process.
      ,
      • Razurel C.
      • Kaiser B.
      • Antonietti J.P.
      • Epiney M.
      • Sellenet C.
      Relationship between perceived perinatal stress and depressive symptoms, anxiety, and parental self-efficacy in primiparous mothers and the role of social support.
      ,
      • Schwab-Reese L.M.
      • Schafer E.J.
      • Ashida S.
      Associations of social support and stress with postpartum maternal mental health symptoms: Main effects, moderation, and mediation.
      ). The stress process model, however, does not take into account multidirectional explanations of depressive symptoms or what happens when mediation by perceived social support is minimal. For instance,
      • Schwab-Reese L.M.
      • Schafer E.J.
      • Ashida S.
      Associations of social support and stress with postpartum maternal mental health symptoms: Main effects, moderation, and mediation.
      found that perceived social support and perceived stress were associated with depressive symptoms during the postpartum period; however, mediation by perceived social support was minimal. This finding suggests that there may be other potential explanations or mechanisms that may better explain the interrelations between these factors and postpartum depressive symptoms. For example, an alternative explanation could be that poor perceived social support predicts perceived stress, and in turn explains postpartum depressive symptoms. Researchers have suggested that low perceived social support is a source of postpartum perceived stress (
      • Andersson E.
      • Hildingsson I.
      Mother’s postnatal stress: An investigation of links to various factors during pregnancy and post-partum.
      ,
      • Cornish D.L.
      • Dobie S.R.
      Social support in the “fourth trimester”: A qualitative analysis of women at 1 month and 3 months postpartum.
      ), and that this increase in perceived stress can increase the risk of depressive symptoms. To this end, our study tested the alternative hypothesis that perceived stress rather than perceived social support mediates the association between poor perceived social support and postpartum depressive symptoms up to 24 months postpartum (Figure 1).
      Figure thumbnail gr1
      Figure 1Mediation conceptual framework. Examining the 6-, 12-, and 18-month postpartum stress as mediators between the 1-, 6-, and 12-month postpartum social support and 12-, 18-, and 24-month postpartum depressive symptoms. a = pathway between predictor and mediator; b = pathway between mediator and outcome; c1 = pathway between predictor and outcome (direct effect).
      Furthermore, although clinicians strive to provide optimal support for postpartum women, clinical care during the postpartum period is generally limited to a single visit at 4 to 6 weeks (unless there are considerable issues after birth requiring more intensive care). This decreases the chances that women will be diagnosed with postpartum depression beyond 4 and 6 weeks after birth (
      • American College of Obstetricians and Gynecologists
      Optimizing postpartum care.
      ). Subsequently, symptoms of depression after this visit can often go underdiagnosed or undertreated (
      • Becker M.
      • Weinberger T.
      • Chandy A.
      • Schmukler S.
      Depression during pregnancy and postpartum.
      ). To this end, researchers and clinicians have called for an extension of the time period for the clinical criteria (i.e., 6 months postpartum) given the research evidence showing that symptoms of depression are present long after 4 weeks after birth (
      • O’Hara M.W.
      • McCabe J.E.
      Postpartum depression: Current status and future directions.
      ). However, research examining longitudinal predictors of depressive symptoms beyond the first few postpartum months is scant. Most studies focus on depressive symptoms before, during, or immediately after pregnancy. Examining the extent to which perceived social support and perceived stress longitudinally predict depression is important because these predictors may change over time and thus increase the risk of depression at certain points during the postpartum period. For example, researchers have shown that perceived instrumental social support (e.g., childcare, buying diapers) decreases throughout the first year of the postpartum period, which can lead to women feeling lonely and experiencing negative psychological well-being (
      • Chen H.H.
      • Hwang F.M.
      • Lin L.J.
      • Han K.C.
      • Lin C.L.
      • Chien L.Y.
      Depression and social support trajectories during 1 year postpartum among marriage-based immigrant mothers in Taiwan.
      ,
      • Negron R.
      • Martin A.
      • Almog
      • Balbierz A.
      • Howell E.A.
      Social support during the postpartum period: Mothers’ views on needs, expectations, and mobilization of support.
      ). Researchers have also found that women who report high perceived stress during the postpartum period do not see any improvement up to 3 years postpartum (
      • Mughal M.K.
      • Giallo R.
      • Arnold P.
      • Benzies K.
      • Kehler H.
      • Bright K.
      • Kingston D.
      Trajectories of maternal stress and anxiety from pregnancy to three years and child development at 3 years of age. Findings from the All Our Families (AOF) pregnancy cohort.
      ). In contrast, researchers have found that some women exhibit a trajectory where depression continuously increases throughout the first year of the postpartum period (
      • Bayrampour H.
      • Tomfohr L.
      • Tough S.
      Trajectories of perinatal depressive and anxiety symptoms in a community cohort.
      ). Nevertheless, changes in these psychosocial factors support the premise that maternal psychosocial well-being may be variable during the postpartum period. Given the discrepancy in the time period for defining postpartum depression and the fact that psychosocial well-being is variable over the course of the postpartum period, there is a need to understand the contributing factors of depressive symptoms beyond the first 4 postpartum weeks.
      The first purpose of this secondary analysis from the First Baby Study (FBS) was to examine change in perceived social support, perceived stress, and depressive symptoms across 1, 6, 12, 18, and 24 months postpartum. The second purpose was to examine whether perceived social support and perceived stress predicted depressive symptoms over time. The third purpose was to explore whether perceived stress mediated the association between perceived social support and depressive symptoms. Based on past evidence (
      • Christenson A.
      • Johansson E.
      • Reynisdottir S.
      • Torgerson J.
      • Hemmingsson E.
      Women’s perceived reasons for their excessive postpartum weight retention: A qualitative interview study.
      ,
      • Ghaedrahmati M.
      • Kazemi A.
      • Kheirabadi G.
      • Ebrahimi A.
      • Bahrami M.
      Postpartum depression risk factors: A narrative review.
      ,
      • Grote N.D.
      • Bledsoe S.E.
      Predicting postpartum depressive symptoms in new mothers: The role of optimism and stress frequency during pregnancy.
      ,
      • Norhayati M.N.
      • Nik Hazlina N.H.
      • Asrenee A.R.
      • Wan Emilin W.M.A.
      Magnitude and risk factors for postpartum symptoms: A literature review.
      ,
      • Reid K.M.
      • Taylor M.G.
      Social support, stress, and maternal postpartum depression: A comparison of supportive relationships.
      ,
      • Robertson E.
      • Grace S.
      • Wallington T.
      • Stewart D.E.
      Antenatal risk factors for postpartum depression: A synthesis of recent literature.
      ,
      • Schwab-Reese L.M.
      • Schafer E.J.
      • Ashida S.
      Associations of social support and stress with postpartum maternal mental health symptoms: Main effects, moderation, and mediation.
      ) and
      • Pearlin L.I.
      • Lieberman M.A.
      • Menaghan E.G.
      • Mullan J.T.
      The stress process.
      urging to consider alternative pathways, we proposed an alternative hypothesis that perceived stress would mediate the association between perceived social support and depressive symptoms over the 24-month postpartum period. We also hypothesized that a) perceived social support would decrease over the 24 months, whereas perceived stress and depressive symptoms would increase, and b) lower perceived social support and higher perceived stress would predict depressive symptoms after accounting for the influence of past depressive symptoms.

      Methods

      Participants

      Women (N = 3,006) were first time mothers from a larger longitudinal study, the FBS (
      • Kjerulff K.H.
      • Velott D.L.
      • Zhu J.
      • Chuang C.H.
      • Hillemeier M.M.
      • Paul I.M.
      • Repke J.T.
      Mode of first delivery and women’s intentions for subsequent childbearing: Findings from the First Baby Study.
      ), examining the association between mode of delivery and subsequent childbearing outcomes. Participants were recruited through flyers at hospitals, childbirth classes, low-income clinics, and ultrasound/community health centers and through local press releases and advertisements. Inclusion criteria were 18 to 35 years old, residents of Pennsylvania, nulliparous, and able to read and speak English or Spanish. A more detailed explanation of the FBS can be found elsewhere (
      • Kjerulff K.H.
      • Velott D.L.
      • Zhu J.
      • Chuang C.H.
      • Hillemeier M.M.
      • Paul I.M.
      • Repke J.T.
      Mode of first delivery and women’s intentions for subsequent childbearing: Findings from the First Baby Study.
      ). Because we were interested in understanding depressive symptoms up to 24 months after delivery and not symptoms from a subsequent pregnancy during this follow-up period, women were excluded from these secondary analyses if they became pregnant again (n = 1,657); of the women who had not become pregnant, women were excluded if they did not report baseline demographics (n = 33). Thus, a final sample of 1,316 women was used for the study analyses.

      Procedures

      This study used data collected from telephone interviews at 1, 6, 12, 18, and 24 months postpartum. Women were surveyed via telephone interviews by trained, professional interviewers employed by the Penn State Harrisburg Center for Survey Research and were asked to respond to questions regarding their experiences with pregnancy and childbearing. The FBS was approved by the Penn State College of Medicine and participating hospital Institutional Review Boards.

      Measures

      Maternal postpartum depressive symptoms were assessed with an adapted version of the Edinburgh Postnatal Depression Scale (EPDS;
      • Cox J.L.
      • Holden J.M.
      • Sagovsky R.
      Detection of postnatal depression: Development of the 10-item Edinburgh Postnatal Depression Scale.
      ). The valid and reliable EPDS is commonly used (
      • Daley A.J.
      • Blamey R.V.
      • Jolly K.
      • Roalfe A.K.
      • Turner K.M.
      • Coleman S.
      • MacArthur C.
      A pragmatic randomized controlled trial to evaluate the effectiveness of a facilitated exercise intervention as a treatment for postnatal depression: The PAM-PeRS trial.
      ,
      • Di Florio A.D.
      • Putnam K.
      • Altemus M.
      • Apter G.
      • Bergink V.
      • Bilszta J.
      • Postpartum Depression: Action Towards Causes and Treatment Consortium
      The impact of education, country, race and Scale.
      ) and includes 10 items (e.g., “I have been so unhappy that I have been crying”) rated on a 4-point response scale ranging from 1 (e.g., Yes, most of the time) to 4 (e.g., No, not at all;
      • Cox J.L.
      • Holden J.M.
      • Sagovsky R.
      Detection of postnatal depression: Development of the 10-item Edinburgh Postnatal Depression Scale.
      ). Based on past evidence from pilot studies (
      • Abbasi S.
      • Chuang C.H.
      • Dagher R.
      • Zhu J.
      • Kjerulff K.
      Unintended pregnancy and postpartum depression among first-time mothers.
      ,
      • Kinsey C.B.
      • Baptiste-Roberts K.
      • Zhu J.
      • Kjerulff K.H.
      Effect of previous miscarriage on depressive symptoms during subsequent pregnancy and postpartum in the First Baby Study.
      ,
      • McCall-Hosenfeld J.S.
      • Phiri K.
      • Schaefer E.
      • Zhu J.
      • Kjerulff K.
      Trajectories of depressive symptoms throughout the peri- and postpartum period: Results from the First Baby Study.
      ), the item “Things have been getting on top of me” was changed to “I have had trouble coping” because women found the term “getting on top of me” confusing. Also, the item “The thought of harming myself has occurred to me” was changed to “The thought of harming myself or others has occurred to me” because postpartum depression has been linked to homicidal ideation, particularly as concerns of harming the infant (
      • Hatters Friedman S.
      • Resnick P.J.
      Child murder by mothers: Patterns and prevention.
      ,
      • Humenik A.L.F.
      • Fingerhut R.
      A pilot study assessing the relationship between child harming thoughts and postpartum depression.
      ,
      • Jennings K.D.
      • Ross S.
      • Popper S.
      • Elmore M.
      Thoughts of harming infants in depressed and nondepressed mothers.
      ). To our knowledge, we are the first researchers to change this item to include homicidal ideation in addition to self-harm. These modified items exhibited good corrected item–total correlations throughout the study. Total scores were calculated, with higher scores indicating greater depressive symptoms. Women were also categorized as no possible depression (EPDS < 10) or possible depression (EPDS ≥ 10), consistent with the literature (
      • Gibson J.
      • McKenzie-McHarg K.
      • Shakespeare J.
      • Price J.
      • Gray R.
      A systematic review of studies validating the Edinburgh Postnatal Depression Scale in antepartum and postpartum women.
      ). It is important to note that this measure is not a clinical diagnostic tool for postpartum depression, but rather a measure of postpartum depressive symptoms. Internal consistencies of the modified EPDS were good at all time points (Cronbach's alpha = 0.80–0.84).
      An adapted version of the Prenatal Psychosocial Profile Hassles Scale was also used to measure perceived stress (
      • Curry M.A.
      • Campbell R.A.
      • Christian M.
      Validity and reliability testing of the prenatal psychosocial profile.
      ,
      • Misra D.
      • O’Campo P.
      • Strobino D.
      Testing a sociomedical model for preterm delivery.
      ). The scale was originally developed for low-income, minority women living in urban areas. Pilot testing in our area indicated that two items exhibited poor item–total correlations in our study population (primarily middle- and upper-income women living in rural and suburban areas) because very few of the respondents reported experiencing those problems. Thus, a modified version of the scale was used: the items “sexual, emotional or physical abuse” and “problems with alcohol or drugs” were changed to “fights with partner” and “fights with other family members.” We chose these new items based on several focus group studies of women during and after pregnancy. These new items exhibited good corrected item–total correlations, as we have reported previously (
      • Kinsey C.B.
      • Baptiste-Roberts K.
      • Zhu J.
      • Kjerulff K.H.
      Effect of previous miscarriage on depressive symptoms during subsequent pregnancy and postpartum in the First Baby Study.
      ,
      • Phelan A.L.
      • DiBenedetto M.R.
      • Paul I.M.
      • Zhu J.
      • Kjerulff K.H.
      Psychosocial stress during first pregnancy predicts infant health outcomes in the first postnatal year.
      ). Women rated 12 items that may have occurred and caused them to feel stressed (i.e., “worries about food, shelter, health care, and transportation,” and “problems with the baby”) using 4-point response options ranging from 1 (no stress) to 4 (severe stress). Total scores were calculated, with higher scores indicating greater perceived stress. Internal consistencies of the adapted scale were acceptable at all time points (Cronbach's alpha = 0.71–0.77).
      Perceived social support was measured using the five-item validated form of the Medical Outcomes Study-Social Support Survey (
      • McCarrier K.
      • Bushnell D.
      • Martin M.
      • Paczkowski R.
      • Nelson D.
      • Buesching D.
      Validation and psychometric evaluation of a 5-item measure of perceived social support.
      ), a commonly utilized shortened scale (e.g.,
      • Ford-Gilboe M.
      • Varcoe C.
      • Scott-Storey K.
      • Wuest J.
      • Case J.
      • Currie L.M.
      • Wathen C.N.
      A tailored online safety and health intervention for women experiencing intimate partner violence: The iCAN plan 4 safety randomized controlled trial protocol.
      ). Women were asked to rate five items (e.g., “someone to confide in or talk to about your problems”) on a 5-point response scale ranging from 1 (none of the time) to 5 (all of the time). Total scores were calculated, with higher scores indicating greater perceived social support. Internal consistencies were good to excellent at all time points (Cronbach's alpha = 0.87–0.92).

      Demographics

      Women reported age, race/ethnicity, family income, marital status, education, current height, and current weight. Prepregnancy weight and height were retrospectively reported to calculate prepregnancy body mass index. Based on past research (
      • Fisher S.D.
      • Wisner K.L.
      • Clark C.T.
      • Sit D.K.
      • Luther J.F.
      • Wisniewski S.
      Factors associated with onset timing, symptoms, and severity of depression identified in the postpartum period.
      ,
      • Palumbo G.
      • Mirabella F.
      • Gigantesco A.
      Positive screening and risk factors for postpartum depression.
      ,
      • Salehi-Pourmehr H.
      • Mohammad-Alizadeh S.
      • Jafarilar-Agdam N.
      • Rafiee S.
      • Farshbaf-Khalili A.
      The association between pre-pregnancy obesity and screening results of depression for all trimesters of pregnancy, postpartum and 1 year after birth: A cohort study.
      ,
      • Silverman M.E.
      • Reichenberg A.
      • Savitz D.A.
      • Cnattingius S.
      • Lichtenstein P.
      • Hultman C.M.
      • Larsson H.
      • Sandin S.
      The risk factors for postpartum depression: A population-based study.
      ), each analysis was run unadjusted and adjusted for the following demographics: weight status, marital status, education, income, age, and race. For example, women who are Caucasian, older, and divorced/separated are at an elevated risk for postpartum depressive symptoms (
      • Fisher S.D.
      • Wisner K.L.
      • Clark C.T.
      • Sit D.K.
      • Luther J.F.
      • Wisniewski S.
      Factors associated with onset timing, symptoms, and severity of depression identified in the postpartum period.
      ,
      • Palumbo G.
      • Mirabella F.
      • Gigantesco A.
      Positive screening and risk factors for postpartum depression.
      ). Further, researchers have found that weight status is associated with postpartum depression such that a higher weight status (i.e., having overweight or obesity) is associated with an increased risk of postpartum depressive symptoms, potentially owing to dysregulation of biological mechanisms (e.g., hypothalamic–pituitary–adrenal axis regulation;
      • Salehi-Pourmehr H.
      • Mohammad-Alizadeh S.
      • Jafarilar-Agdam N.
      • Rafiee S.
      • Farshbaf-Khalili A.
      The association between pre-pregnancy obesity and screening results of depression for all trimesters of pregnancy, postpartum and 1 year after birth: A cohort study.
      ,
      • Silverman M.E.
      • Reichenberg A.
      • Savitz D.A.
      • Cnattingius S.
      • Lichtenstein P.
      • Hultman C.M.
      • Larsson H.
      • Sandin S.
      The risk factors for postpartum depression: A population-based study.
      ).

      Data Analysis

      All data were analyzed with IBM SPSS Statistics (version 24.0). Means, standard deviations, and percentages were used to examine demographic variables and descriptives of primary study variables across the total sample controlling for weight status, marital status, education, income, age, and race. Partial correlations also controlling for these demographic factors were conducted to examine associations between all study variables across the 24 months. Repeated measures analysis of variance (ANOVA) examined changes in primary study variables across the 24 months. Linear regression determined whether perceived social support and perceived stress were predictors of postpartum depressive symptoms. Regression models included perceived social support or perceived stress from the prior time period while controlling for past depressive symptoms. Finally,
      • Hayes A.F.
      Introduction to mediation, moderation, and conditional process analysis: A regression-based approach.
      methods using the PROCESS command in SPSS examined whether postpartum perceived stress mediated the associations between perceived social support and depressive symptoms. Mediation models examined each pathway prospectively (i.e., independent variable at T1, mediator at T2, and dependent variable at T3), which resulted in three parallel models covering unique 12-month spans (Figure 1): 1) 1-month perceived social support (predictor) and 6-month perceived stress (mediator) predicting 12-month depressive symptoms (outcome), 2) 6-month perceived social support (predictor) and 12-month perceived stress (mediator) predicting 18-month depressive symptoms (outcome), and 3) 12-month perceived social support (predictor) and 18-month perceived stress (mediator) predicting 24-month depressive symptoms (outcome). These time points were chosen because they correspond with when the data were collected for the FBS. A significant indirect effect (i.e., predictor predicts the outcome through the mediator) meant the mediator significantly mediated the association between the predictor and outcome.
      • Hayes A.F.
      Introduction to mediation, moderation, and conditional process analysis: A regression-based approach.
      macro features bias-corrected bootstrapping, which determines the effect size of the indirect effect (i.e., β) alongside a 95% confidence interval (i.e., significant indirect effect sizes signaled by p values of <.05 and confidence intervals that do not include zero). Repeated measures ANOVAs, linear regressions, and mediation models analyses adjusted for weight status, marital status, education, income, age, and race did not differ from unadjusted models; thus, only the unadjusted analyses are presented in detail. Repeated measures ANOVAs omit observations listwise, therefore if any observations over time were missing for a participant, that participant was removed. For perceived social support, 3 women were excluded; for perceived stress, 9 women were excluded; and for depressive symptoms, 10 women were excluded. For the linear regression and mediation models, we included only women with no missing data for perceived social support, perceived stress, and perceived depressive symptoms at all time points (N = 1,301). There were no significant differences in study variables between women who did and did not complete all measures.

      Results

      The mean prepregnancy body mass index was 26.19 (SD = 5.96; range, 14.14–45.52) and mean age was 27.6 years old (SD = 4.2); the majority of participants were married or partnered (90.0%), White (86.9%), had a college degree (58.4%), and reported a family income of $75,000 or more (46.1%; Table 1). We identified an increase in the frequency with which women reported scale values that indicated possible depression (i.e., EPDS ≥ 10;
      • Gibson J.
      • McKenzie-McHarg K.
      • Shakespeare J.
      • Price J.
      • Gray R.
      A systematic review of studies validating the Edinburgh Postnatal Depression Scale in antepartum and postpartum women.
      ) at 1 (9.5%), 6 (8.3%), 12 (10.2%), 18 (10.1%), and 24 months (12.1%) with significant increases from 6 to 12 months and 18 to 24 months (McNemar test p < .05).
      Table 1Demographic Characteristics of the Study Sample
      VariableMeanSDPercentage
      Prepregnancy BMI26.195.96
      Age27.564.20
      Marital status
       Married/partnered90.0
       Other10.0
      Race/ethnicity
       White86.9
       Other13.1
      Education
       ≤ High school41.6
       ≥ College58.4
      Family income
       <$20,0009.7
       $20,000–$49,99921.6
       $50,000–$74,99922.6
       ≥$75,00046.1
      Abbreviations: BMI, body mass index; SD, standard deviation.

      Change Over Time and Associations

      Means, standard deviations, and correlations for perceived social support, perceived stress, and depressive symptoms at each time point are presented in Table 2. Figure 2 illustrates the means over time. Repeated measures ANOVAs revealed a significant multivariate effect of time for perceived social support, Wilks’ lambda (Λ) = 0.89, F(4, 1309) = 39.97, p < .001, ηp2 = 0.11, perceived stress, Λ = 0.69, F(4, 1301) = 149.08, p < .001, ηp2 = 0.31, and depressive symptoms, Λ = 0.98, F(4, 1302) = 8.02, p < .001, ηp2 = 0.02. Perceived social support decreased (p < .01) over the 24 months, except from 12 to 18 months postpartum, in which perceived social support remained constant. Perceived stress increased (p < .01) at each time point. Depressive symptoms remained constant until 18 months postpartum and then increased at 24 months (p < .01).
      Table 2Descriptives and Partial Correlations Controlling for Maternal Age, Marital Status, Education, Race, and Family Income Among All Study Variables across 24 Months Postpartum
      Variables1 Month6 Months12 Months18 Months24 Months
      SSStDSSSStDSSSStDSSSStDSSSStDS
      1 Month
       SS−0.34
      p < .01.
      −0.32
      p < .01.
      0.61
      p < .01.
      −0.29
      p < .01.
      −0.25
      p < .01.
      0.58
      p < .01.
      −0.26
      p < .01.
      −0.24
      p < .01.
      0.56
      p < .01.
      −0.25
      p < .01.
      −0.24
      p < .01.
      0.56
      p < .01.
      −0.22
      p < .01.
      −0.20
      p < .01.
       St0.44
      p < .01.
      −0.31
      p < .01.
      0.60
      p < .01.
      0.40
      p < .01.
      −0.31
      p < .01.
      0.54
      p < .01.
      0.35
      p < .01.
      −0.31
      p < .01.
      0.48
      p < .01.
      0.35
      p < .01.
      −0.28
      p < .01.
      0.45
      p < .01.
      0.30
      p < .01.
       DS−0.29
      p < .01.
      0.34
      p < .01.
      0.52
      p < .01.
      −0.28
      p < .01.
      0.29
      p < .01.
      0.44
      p < .01.
      −0.29
      p < .01.
      0.21
      p < .01.
      0.37
      p < .01.
      −0.25
      p < .01.
      0.22
      p < .01.
      0.36
      p < .01.
      6 Months
       SS−0.39
      p < .01.
      −0.39
      p < .01.
      0.68
      p < .01.
      −0.35
      p < .01.
      −0.33
      p < .01.
      0.67
      p < .01.
      −0.34
      p < .01.
      −0.34
      p < .01.
      0.64
      p < .01.
      −0.29
      p < .01.
      −0.32
      p < .01.
       St0.59
      p < .01.
      −0.33
      p < .01.
      0.70
      p < .01.
      0.47
      p < .01.
      −0.38
      p < .01.
      0.65
      p < .01.
      0.47
      p < .01.
      −0.36
      p < .01.
      0.63
      p < .01.
      0.46
      p < .01.
       DS−0.29
      p < .01.
      0.45
      p < .01.
      0.62
      p < .01.
      −0.33
      p < .01.
      0.39
      p < .01.
      0.57
      p < .01.
      −0.30
      p < .01.
      0.40
      p < .01.
      0.55
      p < .01.
      12 Months
       SS−0.40
      p < .01.
      −0.40
      p < .01.
      0.69
      p < .01.
      −0.32
      p < .01.
      −0.32
      p < .01.
      0.67
      p < .01.
      −0.29
      p < .01.
      −0.29
      p < .01.
       St0.56
      p < .01.
      −0.38
      p < .01.
      0.70
      p < .01.
      0.47
      p < .01.
      −0.33
      p < .01.
      0.64
      p < .01.
      0.43
      p < .01.
       DS−0.35
      p < .01.
      0.46
      p < .01.
      0.63
      p < .01.
      −0.30
      p < .01.
      0.39
      p < .01.
      0.56
      p < .01.
      18 Months
       SS−0.39
      p < .01.
      −0.39
      p < .01.
      0.72
      p < .01.
      −0.33
      p < .01.
      −0.32
      p < .01.
       St0.59
      p < .01.
      −0.37
      p < .01.
      0.69
      p < .01.
      0.48
      p < .01.
       DS−0.35
      p < .01.
      0.49
      p < .01.
      0.63
      p < .01.
      24 Months
       SS−0.40
      p < .01.
      −0.40
      p < .01.
       St0.59
      p < .01.
       DS
      Mean21.8715.504.4021.5516.884.1421.2317.364.3421.1617.644.4120.7318.014.71
      SD3.383.123.713.403.793.723.494.093.783.584.293.923.834.184.11
      Abbreviations: DS, depressive symptoms; SS, social support; St, stress.
      The mean values of the constructs are visually represented in Figure 2.
      p < .01.
      Figure thumbnail gr2
      Figure 2Mean scores for perceived social support, perceived stress, and depressive symptoms over the 24-month postpartum period. Note: Perceived social support, n = 1313; perceived stress, n = 1305; and depressive symptoms, n = 1306. Possible scale ranges varied for social support (5–25), stress (12–48), and depressive symptoms (0–30). EPDS, Edinburgh Postnatal Depression Scale; MOS-SSS, Medical Outcomes Study-Social Support Survey. Stress total score was measured via the Prenatal Psychosocial Profile Hassles Scale.
      Pearson correlation analyses (Table 2) revealed significant associations between all study variables at all time points, such that lower levels of perceived social support were associated with higher levels of perceived stress and depressive symptoms whereas higher levels of perceived stress were associated with higher levels of depressive symptoms.

      Regression

      The final model including 1-month depressive symptoms, perceived social support, and perceived stress explained 31% of the variance in depressive symptoms at 6 months postpartum (Table 3). Higher levels of depressive symptoms (p < .001) and perceived stress (p < .001) at 1 month were significant predictors of depressive symptoms at 6 months postpartum. Perceived social support at 1 month trended toward predicting depressive symptoms at 6 months postpartum (p = .07).
      Table 3Predicting Postpartum Depressive Symptoms Using Perceived Social Support and Perceived Stress (N = 1,301)
      Regression ModelsR2Fdfpβ1β2
      Predicting 6-month postpartum depressive symptoms
       Block 1: past postpartum depressive symptoms0.27488.321, 1,300<.001
      1 month0.52
      p < .001.
      0.42
      p < .001.
       Block 20.31193.273, 1,300<.001
      1-month social support−0.05
      1-month stress0.19
      p < .001.
      Predicting 12-month postpartum depressive symptoms
       Block 1: past postpartum depressive symptoms0.41442.612, 1,300<.001
      1 month0.16
      p < .001.
      0.15
      p < .001.
      6 months0.54
      p < .001.
      0.44
      p < .001.
       Block 20.42235.904, 1,300<.001
      6-month social support−0.06
      p < .01.
      6-month stress0.13
      p < .001.
      Predicting 18-month postpartum depressive symptoms
       Block 1: past postpartum depressive symptoms0.45350.943, 1,300<.001
      1 month0.030.03
      6 months0.27
      p < .001.
      0.25
      p < .001.
      12 months0.45
      p < .001.
      0.38
      p < .001.
       Block 20.46220.555, 1,300<.001
      12-month social support−0.04
      12-month stress0.12
      p < .001.
      Predicting 24-month postpartum depressive symptoms
       Block 1: past postpartum depressive symptoms0.47285.294, 1,300<.001
      1 month0.030.04
      6 months0.20
      p < .001.
      0.20
      p < .001.
      12 months0.16
      p < .001.
      0.14
      p < .001.
      18 months0.41
      p < .001.
      0.34
      p < .001.
       Block 20.48197.166, 1,300<.001
      18-month social support−0.02
      18-month stress0.12
      p < .001.
      Abbreviations: df, degrees of freedom; β, standardized regression coefficients.
      p < .001.
      p < .01.
      The final model including 1- and 6-month depressive symptoms and 6-month perceived social support and perceived stress explained 42% of the variance in depressive symptoms at 12 months postpartum. Higher levels of depressive symptoms at 1 and 6 months (p < .001), higher levels of perceived stress (p < .001), and lower levels of perceived social support (p < .05) at 6 months were significant predictors of depressive symptoms at 12 months postpartum (p < .001).
      The final model including 1-, 6-, and 12-month depressive symptoms and 12-month perceived social support and perceived stress explained 46% of the variance in depressive symptoms at 18 months postpartum. Higher levels of depressive symptoms at 6 and 12 months (p < .001) and higher levels of perceived stress at 12 months were significant predictors of depressive symptoms at 18 months postpartum (p < .001). Perceived social support at 12 months was not a significant predictor of depressive symptoms at 18 months postpartum (p > .05).
      The final model including 1-, 6-, 12-, and 18-month depressive symptoms and 18-month perceived social support and perceived stress explained 48% of the variance in depressive symptoms at 24 months postpartum. Higher levels of depressive symptoms at 6, 12, and 18 months (p < .001) and higher levels of perceived stress at 18 months were significant predictors of depressive symptoms at 24 months postpartum (p < .001). Perceived social support at 18 months was not a significant predictor of depressive symptoms at 24 months postpartum (p > .05).

      Mediation

      Table 4 summarizes the mediation models, including 1) regression statistics for each overall model (i.e., the prediction of depressive symptoms using perceived social support as a predictor and perceived stress as the mediator), 2) the total indirect effect with mediator included, and 3) 95% confidence intervals of the indirect effect. All mediation models using postpartum perceived stress as a mediator and perceived social support significantly explained 23% to 25% of the variance in depressive symptoms. At all postpartum time points, perceived stress significantly mediated the association between perceived social support and depressive symptoms such that lower levels of perceived social support predicted perceived stress, which in turn predicted depressive symptoms.
      Table 4Mediation Models Predicting Postpartum Depressive Symptoms (Dependent Variable) Using Social Support (Predictor Variable) and Stress (Mediator), N = 1,301
      Mediation ModelsOverall ModelDirect EffectIndirect Effect
      R2Fp Valueβ (SE)95% CIβ (SE)95% CI
      Model 1: 1-month social support and 6-month stress predicting 12-month postpartum depressive symptoms0.23194.01<0.001−0.13 (0.03)[−0.19 to −0.08]−0.14 (0.02)[−0.18 to −0.11]
      Model 2: 6-month social support and 12-month stress predicting 18-month postpartum depressive symptoms0.25220.30<0.001−0.23 (0.03)[−0.29 to −0.17]−0.16 (0.02)[−0.19 to −0.13]
      Model 3: 12-month social support and 18-month stress predicting 24-month postpartum depressive symptoms0.25213.00<0.001−0.18 (0.03)[−0.24 to −0.12]−0.16 (0.02)[−0.20 to −0.13]
      Abbreviations: β, standardized regression coefficients; CI, confidence interval; SE, standard error.
      Degrees of freedom for all overall models = (2, 1,298), p < .001.

      Discussion

      This study aimed to examine an alternative hypothesis that perceived stress would mediate the association between perceived social support and postpartum depressive symptoms up to 24 months after childbirth. Also, we examined changes in perceived social support, perceived stress, and depressive symptoms up to 24 months postpartum. Overall, perceived social support decreased over time, whereas perceived stress increased; depressive symptoms remained constant and then increased from 18 to 24 months postpartum. Lower levels of perceived social support predicted depressive symptoms at 12 months postpartum, whereas perceived stress predicted depressive symptoms at all time points. Further, the association between perceived social support and depressive symptoms at all time points was mediated by perceived stress such that lower levels of perceived social support predicted perceived stress, in turn predicting depressive symptoms. These findings are of particular interest given that the mediation analyses were entirely prospective with the effects consistent at each time period, suggesting that during the postpartum period perceived stress might play an important underlying role in the association between poor perceived social support and depressive symptoms.
      Partial support was found for the hypothesis that perceived social support would decrease, whereas perceived stress and depressive symptoms would increase over the 24 months. Perceived social support decreased from 1 to 24 months postpartum, remaining constant from 12 to 18 months. Perceived stress increased from 1 to 24 months postpartum, whereas depressive symptoms remained constant, then increased at 18 to 24 months. This finding is consistent with past researchers' conclusions that perceived social support declines and perceived stress increases after birth as women often perceive a shift in focus from their needs to their baby's (
      • Andersson E.
      • Hildingsson I.
      Mother’s postnatal stress: An investigation of links to various factors during pregnancy and post-partum.
      ,
      • Christenson A.
      • Johansson E.
      • Reynisdottir S.
      • Torgerson J.
      • Hemmingsson E.
      Women’s perceived reasons for their excessive postpartum weight retention: A qualitative interview study.
      ,
      • Cornish D.L.
      • Dobie S.R.
      Social support in the “fourth trimester”: A qualitative analysis of women at 1 month and 3 months postpartum.
      ). Women also feel that significant others (e.g., spouse, family members, friends) lack understanding of how much support they really need (
      • Razurel C.
      • Bruchon-Schweitzer M.
      • Dupanloup A.
      • Irion O.
      • Epiney M.
      Stressful events, social support and coping strategies of primiparous women during the postpartum period: A qualitative study.
      ). For example, although there is often a lot of support after the birth of the baby, this attention quickly declines within the first few postpartum weeks; many women feel that others do not understand this support is still needed in the continued months after childbirth as women adjust to their role as a mother along with sleepless nights, hormonal changes, and returning to work. Even 1 year after childbirth, women have reported that social support during the postpartum period, particularly from family members, is expected and that they should not have to ask for it (
      • Negron R.
      • Martin A.
      • Almog
      • Balbierz A.
      • Howell E.A.
      Social support during the postpartum period: Mothers’ views on needs, expectations, and mobilization of support.
      ). Women not perceiving support up to 12 months postpartum have also reported negative experiences with psychological well-being (e.g., stress, depression;
      • Negron R.
      • Martin A.
      • Almog
      • Balbierz A.
      • Howell E.A.
      Social support during the postpartum period: Mothers’ views on needs, expectations, and mobilization of support.
      ). Nevertheless, these findings support that perceived social support and perceived stress can vary over the course of the postpartum period.
      Depressive symptoms were generally stable until around 18 months postpartum, and then slightly increased from 18 to 24 months postpartum. This finding is consistent with past research (
      • McCall-Hosenfeld J.S.
      • Phiri K.
      • Schaefer E.
      • Zhu J.
      • Kjerulff K.
      Trajectories of depressive symptoms throughout the peri- and postpartum period: Results from the First Baby Study.
      ,
      • Mora P.A.
      • Bennett I.M.
      • Elo I.T.
      • Mathew L.
      • Coyne J.C.
      • Culhane J.F.
      Distinct trajectories of perinatal depressive symptomology: Evidence from growth mixture modeling.
      ) suggesting that women may have a trajectory of depressive symptoms that stabilizes during the early postpartum period (<12 months) and increases out to 25 months.
      • Mora P.A.
      • Bennett I.M.
      • Elo I.T.
      • Mathew L.
      • Coyne J.C.
      • Culhane J.F.
      Distinct trajectories of perinatal depressive symptomology: Evidence from growth mixture modeling.
      suggested that the risk of having this trajectory of a “late” increase was partly owing to perceived stress. The increase in perceived stress across 24 months postpartum may have contributed to the increase in depressive symptoms later in the postpartum period. Although the overall depressive symptoms scores were low, the proportion of women with possible depression at each time point was consistent with other estimates of postpartum depression prevalence in the 11% to 17% range (
      • Shorey S.
      • Chee C.Y.I.
      • Ng E.D.
      • Chan Y.H.
      • Tam W.W.S.
      • Chong Y.S.
      Prevalence and incidence of postpartum depression among healthy mothers: A systematic review and meta-analysis.
      ,
      • Woody C.A.
      • Ferrari A.J.
      • Siskind D.J.
      • Whiteford
      • Harris M.G.
      A systematic review and meta-regression of the prevalence and incidence of perinatal depression.
      ). Although it is important to note that the majority of women were not meeting the clinical definition for postpartum depression, the experience of some type of symptoms long after childbirth may be problematic owing to additional stressors in the months after birth, such as changing infant needs (e.g., transition to solid foods, crawling, changing sleep patterns) that occur while women are managing their own sources of stress (e.g., trying to lose weight, returning to work, thinking about having another baby;
      • Miller R.L.
      • Pallant J.F.
      • Negri L.M.
      Anxiety and stress in the postpartum: Is there more to postnatal distress than depression?.
      ,
      • Razurel C.
      • Bruchon-Schweitzer M.
      • Dupanloup A.
      • Irion O.
      • Epiney M.
      Stressful events, social support and coping strategies of primiparous women during the postpartum period: A qualitative study.
      ). Nevertheless, these findings show that, regardless of onset and origin, depressive symptoms are still experienced up to 2 years after childbirth, highlighting a need for long-term screening during women's childbearing years and illustrating a gap in the literature in identifying sources of depressive symptoms in the early years after childbirth that may set the stage for long-term depression.
      Partial support was found for the hypothesis that lower levels of perceived social support and higher levels of perceived stress would explain depressive symptoms after considering the role of past depressive symptoms. Past depressive symptoms explained 31% to 48% of the variance in depressive symptoms, which is consistent with past research suggesting the strongest predictor of depressive symptoms is past depressive symptoms (
      • McCall-Hosenfeld J.S.
      • Phiri K.
      • Schaefer E.
      • Zhu J.
      • Kjerulff K.
      Trajectories of depressive symptoms throughout the peri- and postpartum period: Results from the First Baby Study.
      ). Further, perceived social support and perceived stress provided an additional contribution such that lower levels of 6-month perceived social support predicted 12-month depressive symptoms and higher levels of perceived stress predicted depressive symptoms at all time points. The lack of significance with perceived social support predicting depressive symptoms at 18 and 24 months postpartum may be explained by the decreasing levels of perceived social support during this time. Over time, women may have adapted to poor perceived social support, subsequently mitigating the association between poor perceived social support and depressive symptoms. The finding that perceived stress predicted depressive symptoms at all time points is consistent with past researchers’ findings showing a strong association between perceived stress and depression (
      • Ghaedrahmati M.
      • Kazemi A.
      • Kheirabadi G.
      • Ebrahimi A.
      • Bahrami M.
      Postpartum depression risk factors: A narrative review.
      ,
      • Grote N.D.
      • Bledsoe S.E.
      Predicting postpartum depressive symptoms in new mothers: The role of optimism and stress frequency during pregnancy.
      ,
      • Norhayati M.N.
      • Nik Hazlina N.H.
      • Asrenee A.R.
      • Wan Emilin W.M.A.
      Magnitude and risk factors for postpartum symptoms: A literature review.
      ,
      • Robertson E.
      • Grace S.
      • Wallington T.
      • Stewart D.E.
      Antenatal risk factors for postpartum depression: A synthesis of recent literature.
      ,
      • Schwab-Reese L.M.
      • Schafer E.J.
      • Ashida S.
      Associations of social support and stress with postpartum maternal mental health symptoms: Main effects, moderation, and mediation.
      ).
      Further, as predicted and consistent with past research (
      • Ghaedrahmati M.
      • Kazemi A.
      • Kheirabadi G.
      • Ebrahimi A.
      • Bahrami M.
      Postpartum depression risk factors: A narrative review.
      ,
      • Grote N.D.
      • Bledsoe S.E.
      Predicting postpartum depressive symptoms in new mothers: The role of optimism and stress frequency during pregnancy.
      ,
      • Norhayati M.N.
      • Nik Hazlina N.H.
      • Asrenee A.R.
      • Wan Emilin W.M.A.
      Magnitude and risk factors for postpartum symptoms: A literature review.
      ,
      • Reid K.M.
      • Taylor M.G.
      Social support, stress, and maternal postpartum depression: A comparison of supportive relationships.
      ,
      • Robertson E.
      • Grace S.
      • Wallington T.
      • Stewart D.E.
      Antenatal risk factors for postpartum depression: A synthesis of recent literature.
      ,
      • Schwab-Reese L.M.
      • Schafer E.J.
      • Ashida S.
      Associations of social support and stress with postpartum maternal mental health symptoms: Main effects, moderation, and mediation.
      ), support was found for the hypothesis that perceived stress would mediate the association between lower levels of perceived social support and depressive symptoms. Lower levels of perceived social support predicted perceived stress, in turn predicting depressive symptoms. It may be that when women feel a lack of attention toward their needs as a new mother, this increases the risk of perceived stress (
      • Christenson A.
      • Johansson E.
      • Reynisdottir S.
      • Torgerson J.
      • Hemmingsson E.
      Women’s perceived reasons for their excessive postpartum weight retention: A qualitative interview study.
      ,
      • Cornish D.L.
      • Dobie S.R.
      Social support in the “fourth trimester”: A qualitative analysis of women at 1 month and 3 months postpartum.
      ). With perceived stress as an established independent predictor of depressive symptoms, this increase in perceived stress predicts depressive symptoms. Overall, these findings suggest women may experience depressive symptoms as far out as 24 months after childbirth and highlight the important mechanistic role that perceived stress plays in understanding the association between poor perceived social support and depressive symptoms.
      This study is among the first to prospectively examine mediating influence of perceived stress on perceived social support and depressive symptoms in women beyond the early postpartum period. Strengths include a large cohort study, prospective study design, and the ability to conduct longitudinal mediation analyses. This study also contributes to the literature by providing support for alternative explanations to the stress process model and illustrating the need to better understand the complex interrelations between perceived social support, perceived stress, and depressive symptoms. However, some limitations include generalizability of the study findings (e.g., sample was representative of women in Central Pennsylvania who were White, highly educated, and married). These analyses should be replicated across different cultures to address cultural differences in psychosocial well-being during the first 2 years of the postpartum period (e.g., culture can affect reception of social support and can affect the understanding of what the responsibilities of different family members are;
      • Kim H.S.
      • Sherman D.K.
      • Taylor S.E.
      Culture and social support.
      ,
      • Lawley K.A.
      • Willett Z.Z.
      • Scollon C.N.
      • Lehman B.J.
      Did you really need to ask? Cultural variation in emotional responses to providing solicited social support.
      ). Another limitation was the inherent bias from self-report based surveys. Overall, low scores of depressive symptoms could be due to under-reporting, which our study was unable to assess. Also, the self-reported measure of perceived stress included symptoms associated with emotional distress, which could be confounded with depressive symptoms. Similarly, the measure of perceived social support was not able to disentangle the type or source of social support received. Future studies should further investigate the use of alternative assessments of perceived social support, perceived stress, and depressive symptoms to confirm the study findings.

      Implications for Practice and/or Policy

      Overall, these study findings suggest that efforts are needed to screen and potentially treat depressive symptoms beyond the first 6 postpartum weeks given the change in psychosocial factors over the postpartum period, and their potential accumulating effects on maternal psychosocial well-being during a time period that is accompanied by unavoidable stressors. Further, there is a need to implement early intervention strategies aimed at improving perceived social support to manage perceived stress and depressive symptoms during the postpartum period in an effort to prevent a long-term trajectory of depressive symptoms. For example, strengthening support systems may be a useful strategy to manage perceived stress, and in turn reduce depressive symptoms. Implementing strategies in early postpartum months may help to identify ways to maintain support systems longer after birth and improve perceived stress as a means to reduce depressive symptoms. However, it is important to note that explicit recommendations for how to support women throughout 24 months postpartum cannot be made given that our study did not examine other potential variables that might be associated with perceived social support, perceived stress, and depressive symptoms. For example, changes in employment, relationship quality and satisfaction, and competing demands (e.g., childcare) may also help to explain the associations between these psychosocial variables. Future research is needed to examine these factors as well as other potential influencing correlates to confirm these findings.

      Conclusions

      This novel secondary data analysis provides insight on potential long-term trajectories of depressive symptoms throughout the postpartum period. These findings also suggest an alternative framework to understand the interrelations between postpartum perceived social support, perceived stress, and depressive symptoms. Poor perceived social support predicted perceived stress, in turn predicting depressive symptoms throughout 24 months postpartum. These findings can inform future intervention development and be used to improve counseling content for depressive symptoms during the postpartum period.

      Conflict of Interest

      The authors have no conflicts of interest to disclose.

      Acknowledgments

      The authors thank all the study participants and study staff for collecting data.

      References

        • Abbasi S.
        • Chuang C.H.
        • Dagher R.
        • Zhu J.
        • Kjerulff K.
        Unintended pregnancy and postpartum depression among first-time mothers.
        Journal of Women’s Health. 2013; 22
        • American College of Obstetricians and Gynecologists
        Optimizing postpartum care.
        ACOG Committee Opinion. 2018; 131: e140-e150
        • American Psychiatric Association
        Diagnostic and Statistical Manual of Mental Disorders.
        5th edition. Author, Washington, DC2013
        • Andersson E.
        • Hildingsson I.
        Mother’s postnatal stress: An investigation of links to various factors during pregnancy and post-partum.
        Scandinavian Journal of Caring Sciences. 2016; 30: 782-789
        • Bayrampour H.
        • Tomfohr L.
        • Tough S.
        Trajectories of perinatal depressive and anxiety symptoms in a community cohort.
        Journal of Clinical Psychiatry. 2016; 77: e1467-e1473
        • Becker M.
        • Weinberger T.
        • Chandy A.
        • Schmukler S.
        Depression during pregnancy and postpartum.
        Current Psychiatry Reports. 2016; 18: 32
        • Behrendt H.F.
        • Konrad K.
        • Goecke T.W.
        • Fakhrabadi R.
        • Herpetz-Dahlmann B.
        • Firk C.
        Postnatal mother-to-infant attachment in subclinically depressed mothers: Dyads at risk?.
        Psychopathology. 2016; 49: 269-276
        • Chen H.H.
        • Hwang F.M.
        • Lin L.J.
        • Han K.C.
        • Lin C.L.
        • Chien L.Y.
        Depression and social support trajectories during 1 year postpartum among marriage-based immigrant mothers in Taiwan.
        Archives of Psychiatric Nursing. 2016; 30: 350-355
        • Christenson A.
        • Johansson E.
        • Reynisdottir S.
        • Torgerson J.
        • Hemmingsson E.
        Women’s perceived reasons for their excessive postpartum weight retention: A qualitative interview study.
        PLoS One. 2016; 11: e0167731
        • Cornish D.L.
        • Dobie S.R.
        Social support in the “fourth trimester”: A qualitative analysis of women at 1 month and 3 months postpartum.
        Journal of Perinatal Education. 2018; 27: 233-242
        • Cox J.L.
        • Holden J.M.
        • Sagovsky R.
        Detection of postnatal depression: Development of the 10-item Edinburgh Postnatal Depression Scale.
        British Journal of Psychiatry. 1987; 150: 782-786
        • Curry M.A.
        • Campbell R.A.
        • Christian M.
        Validity and reliability testing of the prenatal psychosocial profile.
        Research in Nursing & Health. 1994; 17: 127-135
        • Daley A.J.
        • Blamey R.V.
        • Jolly K.
        • Roalfe A.K.
        • Turner K.M.
        • Coleman S.
        • MacArthur C.
        A pragmatic randomized controlled trial to evaluate the effectiveness of a facilitated exercise intervention as a treatment for postnatal depression: The PAM-PeRS trial.
        Psychological Medicine. 2015; 45: 2413-2425
        • Di Florio A.D.
        • Putnam K.
        • Altemus M.
        • Apter G.
        • Bergink V.
        • Bilszta J.
        • Postpartum Depression: Action Towards Causes and Treatment Consortium
        The impact of education, country, race and Scale.
        Psychological Medicine. 2016; 21: 1-13
        • Dias C.C.
        • Figueiredo B.
        Breastfeeding and depression: A systematic review of the literature.
        Journal of Affective Disorders. 2015; 171: 142-154
        • Fisher S.D.
        • Wisner K.L.
        • Clark C.T.
        • Sit D.K.
        • Luther J.F.
        • Wisniewski S.
        Factors associated with onset timing, symptoms, and severity of depression identified in the postpartum period.
        Journal of Affective Disorders. 2016; 203: 111-120
        • Ford-Gilboe M.
        • Varcoe C.
        • Scott-Storey K.
        • Wuest J.
        • Case J.
        • Currie L.M.
        • Wathen C.N.
        A tailored online safety and health intervention for women experiencing intimate partner violence: The iCAN plan 4 safety randomized controlled trial protocol.
        BMC Public Health. 2017; 17: 273
        • Ghaedrahmati M.
        • Kazemi A.
        • Kheirabadi G.
        • Ebrahimi A.
        • Bahrami M.
        Postpartum depression risk factors: A narrative review.
        Journal of Education and Health Promotion. 2017; 6: 60
        • Gibson J.
        • McKenzie-McHarg K.
        • Shakespeare J.
        • Price J.
        • Gray R.
        A systematic review of studies validating the Edinburgh Postnatal Depression Scale in antepartum and postpartum women.
        Acta Psychiatrica Scandinavica. 2009; 119: 350-364
        • Grote N.D.
        • Bledsoe S.E.
        Predicting postpartum depressive symptoms in new mothers: The role of optimism and stress frequency during pregnancy.
        Health & Social Work. 2007; 32: 107-118
        • Hatters Friedman S.
        • Resnick P.J.
        Child murder by mothers: Patterns and prevention.
        World Psychiatry. 2007; 6: 137-141
        • Hayes A.F.
        Introduction to mediation, moderation, and conditional process analysis: A regression-based approach.
        Guilford Press, New York, NY2013
        • Humenik A.L.F.
        • Fingerhut R.
        A pilot study assessing the relationship between child harming thoughts and postpartum depression.
        Journal of Clinical Psychology in Medical Settings. 2007; 14: 360-366
        • Jennings K.D.
        • Ross S.
        • Popper S.
        • Elmore M.
        Thoughts of harming infants in depressed and nondepressed mothers.
        Journal of Affective Disorders. 1999; 54: 21-28
        • Kim H.S.
        • Sherman D.K.
        • Taylor S.E.
        Culture and social support.
        American Psychologist. 2008; 63: 518-526
        • Kinsey C.B.
        • Baptiste-Roberts K.
        • Zhu J.
        • Kjerulff K.H.
        Effect of previous miscarriage on depressive symptoms during subsequent pregnancy and postpartum in the First Baby Study.
        Maternal and Child Health Journal. 2014; 19: 391-400
        • Kjerulff K.H.
        • Velott D.L.
        • Zhu J.
        • Chuang C.H.
        • Hillemeier M.M.
        • Paul I.M.
        • Repke J.T.
        Mode of first delivery and women’s intentions for subsequent childbearing: Findings from the First Baby Study.
        Paediatric and Perinatal Epidemiology. 2013; 27: 62-71
        • Lawley K.A.
        • Willett Z.Z.
        • Scollon C.N.
        • Lehman B.J.
        Did you really need to ask? Cultural variation in emotional responses to providing solicited social support.
        PLoS One. 2019; 14: e0219478
        • Li Y.
        • Long Z.
        • Cao D.
        • Cao F.
        Social support and depression across the perinatal period: A longitudinal study.
        Journal of Clinical Nursing. 2017; 26: 2776-2783
        • Mamun A.A.
        • Clavarino A.M.
        • Najman J.M.
        • Williams G.M.
        • O’Callaghan M.J.
        • Bor W.
        Maternal depression and the quality of marital relationship: A 14-year prospective study.
        Journal of Womens Health. 2009; 18: 2023-2031
        • McCall-Hosenfeld J.S.
        • Phiri K.
        • Schaefer E.
        • Zhu J.
        • Kjerulff K.
        Trajectories of depressive symptoms throughout the peri- and postpartum period: Results from the First Baby Study.
        Journal of Womens Health. 2016; 25: 1112-1121
        • McCarrier K.
        • Bushnell D.
        • Martin M.
        • Paczkowski R.
        • Nelson D.
        • Buesching D.
        Validation and psychometric evaluation of a 5-item measure of perceived social support.
        Value in Health. 2011; 14: A148
        • Milgrom J.
        • Hirshler Y.
        • Reece J.
        • Holt C.
        • Gemmill A.W.
        Social support – A protective factor for depressed perinatal women.
        International Journal of Environmental Research and Public Health. 2019; 16: 1426
        • Miller R.L.
        • Pallant J.F.
        • Negri L.M.
        Anxiety and stress in the postpartum: Is there more to postnatal distress than depression?.
        BMC Psychiatry. 2006; 6: 12
        • Misra D.
        • O’Campo P.
        • Strobino D.
        Testing a sociomedical model for preterm delivery.
        Paediatric and Perinatal Epidemiology. 2001; 15: 110-122
        • Mora P.A.
        • Bennett I.M.
        • Elo I.T.
        • Mathew L.
        • Coyne J.C.
        • Culhane J.F.
        Distinct trajectories of perinatal depressive symptomology: Evidence from growth mixture modeling.
        American Journal of Epidemiology. 2008; 169: 24-32
        • Mughal M.K.
        • Giallo R.
        • Arnold P.
        • Benzies K.
        • Kehler H.
        • Bright K.
        • Kingston D.
        Trajectories of maternal stress and anxiety from pregnancy to three years and child development at 3 years of age. Findings from the All Our Families (AOF) pregnancy cohort.
        Journal of Affective Disorders. 2018; 234: 318-326
        • Negron R.
        • Martin A.
        • Almog
        • Balbierz A.
        • Howell E.A.
        Social support during the postpartum period: Mothers’ views on needs, expectations, and mobilization of support.
        Maternal and Child Health Journal. 2013; 17: 616-623
        • Norhayati M.N.
        • Nik Hazlina N.H.
        • Asrenee A.R.
        • Wan Emilin W.M.A.
        Magnitude and risk factors for postpartum symptoms: A literature review.
        Journal of Affective Disorders. 2015; 175: 34-52
        • O’Hara M.W.
        • McCabe J.E.
        Postpartum depression: Current status and future directions.
        Annual Review of Clinical Psychology. 2013; 9: 379-407
        • Palumbo G.
        • Mirabella F.
        • Gigantesco A.
        Positive screening and risk factors for postpartum depression.
        European Psychiatry. 2017; 42: 77-85
        • Pearlin L.I.
        • Lieberman M.A.
        • Menaghan E.G.
        • Mullan J.T.
        The stress process.
        Journal of Health and Social Behavior. 1981; 22: 337-356
        • Phelan A.L.
        • DiBenedetto M.R.
        • Paul I.M.
        • Zhu J.
        • Kjerulff K.H.
        Psychosocial stress during first pregnancy predicts infant health outcomes in the first postnatal year.
        Maternal and Child Health Journal. 2015; 19: 2587-2597
        • Razurel C.
        • Bruchon-Schweitzer M.
        • Dupanloup A.
        • Irion O.
        • Epiney M.
        Stressful events, social support and coping strategies of primiparous women during the postpartum period: A qualitative study.
        Midwifery. 2011; 27: 237-242
        • Razurel C.
        • Kaiser B.
        • Antonietti J.P.
        • Epiney M.
        • Sellenet C.
        Relationship between perceived perinatal stress and depressive symptoms, anxiety, and parental self-efficacy in primiparous mothers and the role of social support.
        Women & Health. 2017; 57: 154-172
        • Reid K.M.
        • Taylor M.G.
        Social support, stress, and maternal postpartum depression: A comparison of supportive relationships.
        Social Science Research. 2015; 54: 246-262
        • Robertson E.
        • Grace S.
        • Wallington T.
        • Stewart D.E.
        Antenatal risk factors for postpartum depression: A synthesis of recent literature.
        General Hospital Psychiatry. 2004; 26: 289-295
        • Salehi-Pourmehr H.
        • Mohammad-Alizadeh S.
        • Jafarilar-Agdam N.
        • Rafiee S.
        • Farshbaf-Khalili A.
        The association between pre-pregnancy obesity and screening results of depression for all trimesters of pregnancy, postpartum and 1 year after birth: A cohort study.
        Journal of Perinatal Medicine. 2018; 46: 87-95
        • Schwab-Reese L.M.
        • Schafer E.J.
        • Ashida S.
        Associations of social support and stress with postpartum maternal mental health symptoms: Main effects, moderation, and mediation.
        Women & Health. 2017; 57: 723-740
        • Silverman M.E.
        • Reichenberg A.
        • Savitz D.A.
        • Cnattingius S.
        • Lichtenstein P.
        • Hultman C.M.
        • Larsson H.
        • Sandin S.
        The risk factors for postpartum depression: A population-based study.
        Depression and Anxiety. 2017; 34: 178-187
        • Shorey S.
        • Chee C.Y.I.
        • Ng E.D.
        • Chan Y.H.
        • Tam W.W.S.
        • Chong Y.S.
        Prevalence and incidence of postpartum depression among healthy mothers: A systematic review and meta-analysis.
        Journal of Psychiatric Research. 2018; 104: 235-248
        • Tambag H.
        • Turan Z.
        • Tolun S.
        • Can R.
        Perceived social support and depression levels of women in the postpartum period in Hatay, Turkey.
        Nigerian Journal of Clinical Practice. 2018; 21: 1525-1530
        • Thurgood S.
        • Avery D.M.
        • Williamson L.
        Postpartum depression (PPD).
        American Journal of Clinical Medicine. 2009; 6: 17-22
        • Woody C.A.
        • Ferrari A.J.
        • Siskind D.J.
        • Whiteford
        • Harris M.G.
        A systematic review and meta-regression of the prevalence and incidence of perinatal depression.
        Journal of Affective Disorders. 2017; 219: 86-92

      Biography

      Krista S. Leonard, MS, is a doctoral student in the Department of Kinesiology at Penn State. She studies predictors of postpartum depression and behavioral strategies (physical activity, healthy eating, self-regulation) to help manage weight gain during pregnancy.
      M. Blair Evans, PhD, is an Assistant Professor in the Kinesiology Department at Penn State University. Blair's interest in groups relates to how social relationships shape the behavior and wellbeing of individuals in varying populations, especially the relationships among peers in small groups.
      Kristen H. Kjerulff, MA, PhD, is Professor, Departments of Public Health Sciences and Obstetrics and Gynecology, Penn State University, College of Medicine. She studies maternity/reproductive health care and common treatments and procedures. She was Principal Investigator of the First Baby Study.
      Danielle Symons Downs, PhD, is a Professor of Kinesiology and Obstetrics/Gynecology and Director of the Exercise Psychology Laboratory at Penn State. She is an expert in women's health, motivation, and designing effective health behavior change interventions (exercise, weight) in women before/during/after pregnancy.