Notes on Stressful impact of depression on early mother–infant relations (Milgrom et al., 2006)
Introduction
The article investigates how post-natal depression (PND) relates to sustained parenting stress and whether conventional CBT for maternal mood improves mother–infant interactions, as well as whether a targeted, short mother–infant intervention can yield additional benefits. The authors draw on three data sources: (1) a longitudinal cohort of postpartum depressed and non-depressed women; (2) a treatment study with 162 depressed women receiving cognitive-behavioural therapy (CBT), routine care, and a comparison group of 162 non-depressed women; and (3) a pilot study of a specialized parent–infant intervention with 22 depressed women. They frame a biopsychosocial model in which the mother’s PND (characterized by sadness, flat affect, and negative thinking) disrupts the essential behavioral and emotional exchanges with the infant, while risk factors (historical vulnerabilities and precipitating events) further impair mother–infant interaction. A central question is whether best-practice treatments for PND should be broadened to address mother–infant interactions, given that most current treatments target maternal mood rather than the interaction itself. The authors note that meta-analytic work on mother–infant interventions showed limited coverage of depressed women and that interventions focusing on sensitivity in mother–infant interactions tend to yield larger effects when manualized and of moderate duration. Three core questions guide the study: (1) how long-lasting are mother–infant relationship difficulties associated with PND? (2) does conventional treatment for maternal mood improve the mother–infant relationship? (3) are brief, targeted mother–infant interventions clinically beneficial in this context?
Methods and Design
The study integrates findings from three sources. First, a 42-month longitudinal follow-up of mothers and infants previously described at earlier time points examined whether parenting stress linked to PND persists beyond the early postnatal period. Recruitment occurred at 3–4 months postnatal, using Edinburgh Postnatal Depression Scale (EPDS) screening, with a cut-off score of 11.5 to identify risk for PND. From 6,500 screened, 789 scored ≥12, 505 accepted psychological assessment, and 378 met DSM-IV depression criteria. Of these, 162 completed baseline measures and were willing to participate in the treatment component. A non-depressed control group of 162 women (EPDS < 12) was recruited in parallel. The depressed participants were randomized to 12 sessions of CBT (n = 40) or routine care with a maternal–child health nurse (n = 32); the remainder received other treatments or referrals and contributed baseline data but are not reported here. The depression severity at intake (BDI) averaged 23.8 for the depressed group and 4.9 for the non-depressed group. The CBT program content is described in detail elsewhere and has a dedicated manual. The second data source comprises the CBT trial itself and a routine care comparator, with assessments at pre-treatment and after 12 weeks. The third source is a pilot of a targeted mother–infant module named Happiness, Understanding, Giving & Sharing (HUGS) delivered after CBT completion to a subset of depressed participants who consented to participate. HUGS is designed to directly enhance the mother–infant interaction by promoting sensitive responsiveness, understanding infant cues, and separating parental issues from infant needs. It draws on Field, Fraiberg, Brazelton, Muir, and related work, and targets interaction directly while acknowledging maternal cognitions and beliefs and the role of family of origin. The HUGS module is described as three sessions and is available in manualized form.
Measures
The primary outcome measure for parenting stress was the Parenting Stress Index (PSI), a 101-item self-report instrument that yields seven child-domain subscales, six parent-domain subscales, two domain scores, and a total PSI score. Higher scores indicate greater parenting stress and dysfunctional mother–infant relations. A total PSI score above 259 is considered indicative of a dysfunctional relationship requiring intervention. The PSI comprises 13 subscales in total, with child-domain subscales including Reinforces Parent, Mood, Acceptability, Adaptability, Demandingness, and Hyperactivity/Distractibility, and parent-domain subscales including Sense of Competence, Attachment to Infant, Restriction of Role, Depression, Relationship with Spouse, Social Isolation, and Health. The Beck Depression Inventory (BDI) provided a clinical rating of depression severity, with the revised form used in this study. Demographic data, childhood trauma exposure, and major life stressors around pregnancy were collected. Labour difficulty was rated on a five-point Likert scale.
Statistical Analysis
Analyses were conducted to minimize Type I error, with means reported alongside 95% confidence intervals. Categorical comparisons used Chi-squared tests and odds ratios (ORs) were derived from logistic regression. Continuous data were analyzed with ANOVA where appropriate. For binary outcomes (dysfunctional vs not), PSI scores > 259 were coded as dysfunctional. To report relationships cleanly, the study emphasizes that the overall pattern across PSI subscales favored depressed women showing higher stress across both child and parent domains. In regression analyses, demographic and psychometric variables were examined as predictors of dysfunctional mother–infant relations. The authors report ORs with 95% CIs and note when CI limits do not include 1. They also report p-values for key treatment effects and regression models, and present the R-squared value for the association between post-treatment depression and PSI scores.
Longitudinal Follow-Up: Persistence of Parenting Stress
The 42-month follow-up showed that elevated parenting stress associated with PND persisted, with scores remaining stable after an initial postnatal decline observed between 3 and 6 months. At 42 months, data were available for 70% of the depressed group (22/31) and 60% of the non-depressed group (26/43). Attrition was random with respect to depressive status. The pattern of higher PSI scores in the depressed group persisted, and the magnitude of the difference between depressed and non-depressed groups remained comparable to earlier time points. Figure I illustrates total PSI scores by infant age (3, 6, 12, and 42 months) for depressed and non-depressed mothers, showing that the depressed group consistently demonstrated higher parenting stress over time.
Baseline Characteristics and Group Differences
At intake, the depressed group exhibited substantially higher parenting stress than the non-depressed controls. The mean BDI score for the depressed group was 23.8 (95% CI: 22.6–25.2) and for non-depressed controls 4.9 (95% CI: 4.3–5.6). In the CBT-treated subgroup, the mean BDI at intake was 23.9 (95% CI: 21.2–26.6). Table I presents pre-treatment PSI subscale scores showing consistently higher scores across the depressed group compared with controls. For example, the Child Domain total was 111.86 (95% CI: 108.51–115.20) in the depressed group versus 90.23 (95% CI: 87.60–92.87) in the non-depressed group, and the Total PSI score was 282.46 (95% CI: 276.1–288.81) versus 204.49 (95% CI: 199.07–209.91) for the respective groups. Across subscales such as Sense of Competence, Attachment to Infant, Depression, and Health, the depressed group similarly showed higher scores, indicating more dysfunctional perceptions and interactions. Notably, more than half of the depressed sample (118/162) had total PSI scores above 259, compared with 4/162 in the non-depressed group (Odds ratio, OR ≈ 113 based on the data; the paper emphasizes a highly significant group difference with χ² = 151, p < 0.001).
Effectiveness of Psychological Treatment on Depression and Parenting Stress
In the randomized study, CBT significantly reduced depressive symptoms, with a mean BDI drop of 7.7 points after treatment (95% CI: 3.5–11.9; p = 0.001). Routine care yielded a smaller, non-significant drop of 2.7 points (95% CI: −5.7–11.1; p = 0.49). However, improvements in depression did not translate into parallel improvements in parenting stress. Post-treatment PSI scores remained elevated relative to non-depressed controls, and there was no significant association between post-treatment BDI and post-treatment PSI within the CBT group (R² = 0.04, p = 0.31). Across the entire depressed sample, there was a modest overall reduction in PSI over 12 weeks of treatment (mean ΔPSI ≈ −19.5 points), but CBT-treated and routine-care groups did not differ significantly by post-treatment PSI; both continued to fall within the dysfunctional range. Figure 2 displays PSI trajectories for control, CBT, and routine care groups, illustrating that CBT did not normalize PSI scores relative to non-depressed controls and that the difference between CBT and routine care persisted at post-treatment.
HUGS Pilot: A Short, Targeted Mother–Infant Intervention
The HUGS module was delivered to 22 depressed women who had completed CBT. At entry, their mean BDI was 24.1 (CI: 20–28.3); after CBT, their mean BDI had fallen to 17.0 (CI: 13.2–20.8). Although there is no contemporaneous control group for the HUGS pilot, the PSI outcomes after 3 weeks of HUGS showed large improvements. The reported reduction in the total PSI score over the short course of HUGS was substantial, with a crude weekly rate of improvement during HUGS calculated at approximately 4.9 points per week, which was more than three times the CBT weekly rate of improvement (about 1.6 points per week). Across PSI subscales, 10 of 13 subscales showed numerically lower scores after HUGS compared with before, though the authors caution that multiple subscale tests inflate the Type I error rate. A paired t-test on the parent-domain total scores showed a significant downward movement after HUGS (p = 0.015). The overall total PSI score declined from 156.2 to 140.5 over 3 weeks, a change deemed statistically significant by the authors. Figure 3 depicts rates of change in total PSI for CBT and HUGS, illustrating a faster improvement during HUGS, and Figure 4 shows subscale changes following the HUGS module, with the majority of subscales showing improvements. The authors acknowledge limitations: there was no contemporaneous control group; allocation and treatment were unblinded; and the pilot design precludes strong causal inferences about the HUGS effect. Nevertheless, the data suggest that direct, targeted early intervention in the mother–infant relationship can be implemented as an adjunct to mood-focused treatment and that some mothers report meaningful psychosocial gains in their sense of parenting competence, attachment, and role restriction after HUGS.
Discussion and Interpretation
The study reinforces the robust link between PND, mother–infant interaction difficulties, and subsequent child development issues, noting enduring effects up to 3.5 years post-partum. Conventional treatments for maternal mood (e.g., CBT) yield only modest reductions in parenting stress as measured by the PSI, and these improvements do not consistently translate into normalized mother–infant relations. This argues against relying on mood improvement alone to safeguard infant development and supports the need to address the mother–infant relationship directly. The HUGS pilot provides preliminary evidence that a short, targeted intervention focused on play, infant cue understanding, and distinguishing infant needs from parental stress can substantially reduce parenting stress, particularly in the parent domain (e.g., sense of competence, attachment, and role restriction). The authors emphasize that improvements in maternal mood do not automatically resolve parenting stress and that dynamic, relationship-focused interventions may offer additional protective benefits for infants. They also discuss the biopsychosocial framework, acknowledging that vulnerability factors such as childhood trauma and precipitating factors like labour and pregnancy stress increase the likelihood of dysfunctional mother–infant relations.
Mechanisms and Theoretical Implications
The authors outline a conceptual mechanism: depressive symptoms (flat affect, preoccupied mood, psychomotor retardation) can impede empathic attunement and responsive engagement, leading to diminished infant regulation and engagement. This fosters a cycle of aversion and non-attunement, potentially entrenching dysfunction. The HUGS module attempts to disrupt this cycle through a three-pronged approach: (1) promoting mother–infant engagement via play and physical contact; (2) teaching mothers to observe and interpret infant signals; and (3) examining parental responses to infant cues, grounded in Muir’s “wait, watch and wonder” framework. The module also integrates awareness of childhood trauma and its impact on vulnerability to PND, thereby acknowledging the broader psychosocial context. The paper links these ideas to classic developmental theories (Brazelton, Winnicott, Field) and contemporary CBT concepts, emphasizing a collaborative stance that integrates mood treatment with interaction-focused strategies.
Practical and Ethical Implications
Clinically, the findings suggest that practitioners should consider adding targeted mother–infant interventions to standard mood-focused treatments for PND, especially when long-term parenting stress and potential infant developmental risks are of concern. The results argue for broadened criteria of “best-practice” treatments to explicitly assess and improve mother–infant interactions, not merely maternal mood. Ethically, delayed or incomplete improvement in mother–infant relations despite mood stabilization could have lasting consequences for infant development, reinforcing the case for integrated, multi-component approaches in postnatal care. The authors call for fully controlled trials of the HUGS module to confirm its efficacy and to determine optimal delivery formats, durations, and populations most likely to benefit.
Limitations and Future Directions
Key limitations include the uncontrolled design of the HUGS pilot, lack of contemporaneous control, and unblinded allocation, which constrain causal inferences about HUGS effectiveness. The authors stress that the HUGS findings require replication in a fully controlled trial. They also acknowledge that the generalizability of the HUGS approach needs testing across diverse settings and populations. Future work should explore longer-term outcomes, potential mediators (e.g., maternal responsiveness) and moderators (e.g., infant temperament, partner support), and cost-effectiveness analyses to guide healthcare policy.
Conclusion
The study demonstrates that PND is associated with long-lasting parenting stress and that conventional mood-focused treatments (CBT) produce only modest improvements in the mother–infant relationship. A brief, targeted mother–infant intervention (HUGS) showed more rapid and substantial reductions in parenting stress within a short time frame, particularly in the parent domain, suggesting a potential enhancement to standard care. The authors propose broader, best-practice criteria for PND treatment that explicitly address mother–infant interactions and advocate for further controlled trials to validate the HUGS approach and its integration with mood-focused therapies. The work underscores the importance of addressing both maternal mood and maternal–infant interactions to optimize infant development and family functioning in the context of postnatal depression.
Acknowledgments and References
The authors acknowledge the memory of Rachael McCarthy and note funding from the Australian Rotary Health Research Fund and The Austin Hospital Medical Research Foundation. The reference list includes foundational works on postnatal depression, mother–infant interaction, CBT, psychosocial risk factors, and prior intervention studies (e.g., Abidin, 1986; Als et al., 2004; Field, 1982, 1997; Brazelton et al., 1974; Murray & Cooper, 1997; O’Hara et al., 2000; Wickberg & Hwang, 1996; Winnicott, 1974; Bakermans-Kranenburg et al., 2003). The study situates itself within a broad literature on PND and infant development and contributes new evidence about the potential added value of a brief, targeted mother–infant module to traditional mood-focused treatments.
- Key quantitative findings and concepts quoted and summarized from the study are included with LaTeX-formatted equations where appropriate, such as the dysfunctional PSI threshold ext{PSI} > 259, ORs with95% CIs ext{OR}=8.13, ext{ 95% CI }[4.61,14.49], and the labour-related OR ext{OR}=1.5, ext{ 95% CI }[1.11,2.03], and p-values p=0.001, ext{ }p=0.49, ext{ }p=0.31, ext{ }p=0.015, along with the PSI change of riangle ext{PSI}_{12 ext{wk}} = -19.5 points and weekly rates during CBT vs HUGS of 4.9 ext{ points/week} vs 1.6 ext{ points/week}. A full set of numbers is embedded in the narrative above to reflect the study’s detailed quantitative reporting.