family-support-protects-maternal-mental-health-across-pregnancy-and-postpartum
Family Support Protects Maternal Mental Health Across Pregnancy and Postpartum

Family Support Protects Maternal Mental Health Across Pregnancy and Postpartum

A longitudinal investigation conducted in Aotearoa New Zealand has demonstrated that perceived social support serves as a sustained protective factor against maternal depression and anxiety throughout pregnancy and the postpartum period, with family support emerging as the single most reliable protective source. Published in Archives of Women’s Mental Health by researchers Mehreen Choudhary, Liesje Donkin, and Amy Kercher, the study tracked pregnant participants over twelve months to evaluate how distinct dimensions of social support interact with maternal psychological wellbeing over time. While international literature has broadly documented that interpersonal assistance protects maternal psychological health, empirical investigations within Aotearoa New Zealand have historically left the precise longitudinal mechanisms of support underexamined. By monitoring women across sequential trimesters and postpartum phases, the research team sought to clarify how perceived support relates to maternal distress, directly comparing the independent contributions of partners, relatives, and friends, while testing whether cumulative stressors and sleep disturbances moderate these longitudinal associations.

The theoretical framework underpinning the investigation synthesises classical psychological models with Indigenous conceptualisations of health to better reflect the sociocultural environment of Aotearoa New Zealand. Within Western psychological literature, the stress-buffering hypothesis formulated by Cohen and Wills, along with Schaefer and colleagues, posits that interpersonal assistance promotes psychological wellbeing both directly and by mitigating the harmful impacts of external stressors. Stress and coping theory, articulated by Lazarus and Folkman and expanded by Biggs and colleagues, frames social support as an indispensable coping resource that alleviates psychological distress when external demands overwhelm an individual’s personal capacity. However, because Western paradigms operate from individualistic foundations, the authors emphasised the necessity of incorporating Māori models of health, including Durie’s Te Whare Tapa Whā, which conceptualises whānau, or extended family networks, as a foundational structural pillar of holistic health rather than merely an external coping resource. Grounding attachment in relational concepts such as tūhono, signifying bonding, and tūhonotanga, reflecting collective connectedness and spiritual interdependence rooted in whakapapa, highlights why the continuity and origin of relational assistance remain profoundly consequential for perinatal wellbeing.

To capture these psychosocial dynamics across the perinatal transition, the authors conducted a secondary analysis of data gathered through the observational Maternal Wellbeing and Exercise Through Pregnancy Study between 2023 and 2024. The study gathered quantitative survey responses across two distinct phases, initiating with an exploratory pilot cohort of 34 participants designed to evaluate practical feasibility before expanding into a larger main cohort comprising 87 individuals recruited via social media notices across community and parenting channels. Eligible participants were pregnant individuals aged 18 years and older who completed repetitive online assessments tracking their psychological functioning and behavioral health. After data cleaning excluded 11 individuals due to ineligibility or missing baseline entries, the final cohort comprised 121 women who entered the project across diverse gestational stages. Baseline demographic profiles indicated a mean maternal age of 31.7 years with a standard deviation of 4.3 years, alongside a median gestational age of 24 weeks with an interquartile range spanning 15 to 32 weeks, representing a sample that was predominantly European or Pākehā, highly educated, mostly married or in de facto relationships, and largely employed on a full-time basis.

Data collection relied on validated self-report screening instruments administered across monthly follow-up schedules to track psychological symptoms and interpersonal perceptions. Depressive symptom severity was measured monthly utilising the Edinburgh Postnatal Depression Scale, a widely validated ten-item instrument capturing experiences across the preceding seven days on a zero to thirty scale, which demonstrated sound internal consistency within the sample at a Cronbach’s alpha of 0.84. Perinatal anxiety was monitored at baseline, three, six, nine, and twelve months using the 31-item Perinatal Anxiety Screening Scale, which measures symptom severity on a range from zero to 93 and exhibited an internal reliability coefficient of 0.93. Perceptions of social support were concurrently evaluated at those same quarterly intervals through the twelve-item Multidimensional Scale of Perceived Social Support, using a seven-point Likert scale to gauge assistance from partners, family members, and friends. The perceived support scale exhibited high baseline internal consistency, yielding an overall Cronbach’s alpha of 0.94, with individual domain subscales demonstrating reliability values ranging between 0.93 and 0.94.

The methodological design also integrated explicit evaluations of potential environmental and physiological moderators, focusing specifically on cumulative life stressors and subjective sleep quality. In the main phase of data collection, stressful life events were documented using a comprehensive checklist encompassing acute and ongoing contextual challenges, including severe financial strain, interpersonal relationship conflict, precarious or unstable housing conditions, and pressing health difficulties, which were summed into an aggregate cumulative stressor count. Maternal sleep quality was tracked through Likert-type questions probing multiple facets of sleep, including retrospective evaluations of overall sleep quality across the preceding week. Because sleep assessment wording differed slightly between the preliminary pilot phase and the subsequent main cohort, responses were harmonised into an operational binary indicator denoting the presence or absence of significant sleep problems. Because stressor inventories were not incorporated into the initial pilot cohort, moderation models focusing on cumulative stress exposure primarily reflected responses supplied by participants enrolled during the subsequent main phase.

To standardise the unbalanced longitudinal observations across gestation and infant development, researchers mapped survey completions into six sequential perinatal categories based on reported gestational timing and infant age. The temporal structure designated Trimester 1 as zero to 13 weeks of gestation, Trimester 2 as 14 to 26 weeks, and Trimester 3 as 27 to 40 weeks, while postpartum responses were classified into early postpartum spanning zero to 12 weeks, mid-postpartum covering 13 to 25 weeks, and late postpartum extending from 26 to 52 weeks. Where missing entries occurred, values were assigned using the nearest valid temporal assessment. Data preparation procedures carried out in IBM SPSS Statistics Version 30.0 applied mean imputation where fewer than twenty percent of instrument items were missing. Evaluation of survey attrition indicated that higher baseline depression, elevated baseline anxiety, and lower formal educational attainment were significantly associated with lower survey completion rates, whereas other baseline demographic variables demonstrated no significant predictive influence over survey completion.

Longitudinal inferential modeling was executed in Jamovi Version 2.6.44 using linear mixed models to account for the nested structure of repeated observations across unbalanced monthly timepoints under the assumption that missing observations occurred at random. Linear mixed models were selected because they incorporate all available repeated measurements without requiring complete datasets across all scheduled assessment points, while random intercepts accounted for within-participant clustering over time. Diagnostic assessments verified that model assumptions were fully satisfied, revealing no detrimental multicollinearity among variables. Although perceived partner support scores displayed negative skewness indicative of a ceiling effect, the distribution did not undermine model stability or computational performance. Relying on methodological power standards established by Maas and Hox, the final analytical sample of 121 women providing repeated observational records across twelve months supplied adequate statistical power to reliably detect small-to-moderate longitudinal fixed effects and moderation interactions.

Descriptive trajectories across perinatal stages documented noticeable fluctuations in psychological distress alongside relative stability in perceived interpersonal support. Depressive symptoms measured by the Edinburgh Postnatal Depression Scale peaked during Trimester 1 before declining to their lowest average levels during late postpartum, whereas anxiety scores on the Perinatal Anxiety Screening Scale reached their zenith during early postpartum before subsiding to lowest levels in late postpartum. Total perceived social support remained stable across pregnancy and postpartum, though friend support exhibited a minor decline during the late postpartum window. Sleep disturbances were most prevalent during Trimester 3 and least frequent during Trimester 2, while average cumulative stressor counts remained comparatively low across all stages. Bivariate Pearson correlations revealed that higher social support was consistently associated with lower depression and anxiety across stages, with family support exhibiting the most stable and frequent inverse correlations with maternal distress from Trimester 2 through late postpartum.

Primary linear mixed models adjusted for perinatal stage revealed that higher average total perceived support was significantly associated with lower depressive symptoms and lower anxiety symptoms, with no omnibus differences in distress observed between developmental stages. When support sources were tested in separate models, higher partner support, family support, and friend support were each individually associated with lower maternal depressive symptoms, while partner support and family support also predicted lower anxiety. However, when all three relational sources were incorporated simultaneously into combined multivariable models, family support was the only source that retained independent statistical significance as a negative predictor of both maternal depression and anxiety. Subsequent models examining interactions between perinatal stage and social support yielded non-significant interaction terms, producing parallel symptom trajectories across time and demonstrating that the protective association of perceived social support operates uniformly across pregnancy trimesters and postpartum periods.

Moderation analyses examining environmental stressors and physiological disruption revealed divergent patterns across maternal depression and anxiety outcomes. Cumulative stressor exposure was positively associated with both depressive and anxiety symptoms, while the interaction between total social support and stressors reached statistical significance for anxiety and approached significance for depression. As illustrated by interaction plots, higher perceived social support attenuated the harmful relationship between cumulative stressors and anxiety symptoms, demonstrating a potent stress-buffering effect under elevated stress conditions. Conversely, while sleep problems were significantly associated with elevated depression and anxiety, the interaction between sleep problems and social support failed to achieve statistical significance. This indicates that sleep disruptions contribute directly to psychological distress rather than modifying the protective buffering capacity of perceived social support, an observation potentially reflecting the fact that the Multidimensional Scale of Perceived Social Support measures emotional reassurance rather than instrumental sleep assistance.

Temporal dynamics were further illuminated through lagged linear mixed models designed to test whether perceived social support predicted subsequent psychological distress at future assessments. Lagged support scores significantly predicted lower depressive symptoms at subsequent timepoints, confirming that interpersonal support contributes prospectively to the reduction of depressive pathology over time. In sharp contrast, lagged social support did not significantly predict subsequent anxiety scores. This temporal dissociation suggests that perinatal depression may reflect entrenched, long-term relational patterns that respond progressively to interpersonal reassurance, whereas perinatal anxiety represents an acute psychological state that fluctuates rapidly in response to immediate environmental demands and acute stressors rather than exhibiting delayed relief from antecedent relational support.

The authors identified several methodological constraints that qualify the generalisability of their conclusions. The study sample was relatively homogeneous, consisting predominantly of highly educated, partnered women identifying as European or Pākehā, resulting in a marked underrepresentation of Māori and Pasifika mothers whose cultural contexts might yield distinct relational dynamics. Furthermore, attrition across the twelve-month monitoring period was non-random, as baseline psychological distress and lower educational attainment predicted participant drop-out; although linear mixed models accommodate missing data under the missing-at-random assumption, unmeasured confounding cannot be ruled out entirely. Measurement constraints also warrant consideration, including ceiling effects that restricted variability in partner support scores, lower measurement frequency for anxiety relative to monthly depression assessments, lack of stressor measurement in the preliminary pilot cohort, and the inability of standardized instruments like the Multidimensional Scale of Perceived Social Support to capture instrumental aid, relationship quality, professional guidance, or the chosen whānau networks central to LGBTQIA+ parents.

Despite these caveats, the empirical findings carry substantial implications for perinatal healthcare delivery and family-centred clinical practice across Aotearoa New Zealand and internationally. The authors argue that maternal health services should routinely incorporate standardised screenings for social support, cumulative stressors, and sleep disturbances to identify vulnerable mothers before psychological distress escalates. Because family support emerged as the most resilient protective buffer across both depression and anxiety, clinical interventions must look beyond partner-centric paradigms and actively facilitate family-inclusive support structures. Furthermore, maternal healthcare pathways must adapt to serve diverse family structures—including single mothers, recent migrants isolated from kinship networks, and LGBTQIA+ parents—while embracing Indigenous relational frameworks such as Te Whare Tapa Whā that recognise whānau as an indispensable, integrated foundation of psychological resilience and maternal wellbeing throughout the perinatal transition.

Subject of Research: Longitudinal associations between perceived social support and perinatal maternal depression and anxiety

Article Title: Longitudinal associations between perceived social support and maternal depression and anxiety across the perinatal period

Article References: Choudhary, M., Donkin, L., & Kercher, A. (2026). Longitudinal associations between perceived social support and maternal depression and anxiety across the perinatal period. Archives of Women’s Mental Health, 29(6), Article 150. https://doi.org/10.1007/s00737-026-01785-0

Image Credits: AI Generated

DOI: 10.1007/s00737-026-01785-0

Keywords: Perinatal mental health, Maternal depression, Perinatal anxiety, Perceived social support, Family support, Partner support, Edinburgh Postnatal Depression Scale, Perinatal Anxiety Screening Scale, Multidimensional Scale of Perceived Social Support, Linear mixed models, Stress buffering hypothesis, Aotearoa New Zealand, Maternal wellbeing, Te Whare Tapa Wha, Sleep problems, Psychosocial stressors, Postpartum mental health, Whanau