Objective The critical importance of a secure mother-infant attachment relationship for long-term physical and mental health of the child is well established. Our study aim was to explore mothers’ subjective experience of the mother-infant relationship after discharge from hospital following neonatal cardiac surgery.
Design Participants were 97 infants who underwent cardiac surgery before the age of 3 months and their mothers. Mothers completed Maternal Postnatal Attachment Scale (MPAS) and Edinburgh Postnatal Depression Scale (EPDS) questionnaires and were interviewed after the infant had been discharged home for 4 weeks. Interviews were analysed using inductive thematic analysis.
Results Mean sores on the MPAS were similar to community norms (84.5 (SD 7.2) vs 84.6 (SD 7), p=0.47). 66/91 mothers interviewed described impacts which encompassed four themes; enhanced emotional ties (n=34, 37%), ‘bonding’ difficulties (n=22, 23%), anxiety and worry (n=17, 19%), and caregiving behaviours (n=10, 11%). Mothers who described bonding difficulties had lower MPAS scores (mean 80.6 (SD 10) vs 85.7 (SD 5.7), p=0.0047), were more likely to have a prenatal diagnosis of the cardiac abnormality (OR 2.6, 95% CI 0.89 to 8.9) and higher EPDS score (9.1 (SD 5.3) vs 6.2 (SD 3.9), p=0.01). Higher EPDS scores were associated with lower MPAS scores (r=−0.44, p=0.0001).
Conclusions Most mothers report a positive relationship with their infant following cardiac surgery but almost a quarter have difficulties forming a strong emotional tie. Clinical care (including prenatal) of the infant with congenital heart disease requiring surgery should include screening, assessment and appropriate referral for early intervention if mothers are struggling to form a bond with their infant.
Statistics from Altmetric.com
What is known about this topic
Infants who have cardiac surgery are at risk for cognitive and behavioural problems.
Early maternal feelings of attachment towards their infant have not been studied in this population.
A secure mother-infant attachment relationship is important for optimal development and long-term physical and mental health.
What the study adds
As a whole group mothers’ attachment feelings towards their infants with congenital heart disease do not differ from community samples when measured using a standardised questionnaire.
When interviewed, mothers reported that cardiac surgery affected their relationship with their infant with almost half describing a positive relationship but 23% reporting ‘bonding’ difficulties.
Maternal report of ‘bonding’ difficulties was associated with prenatal diagnosis of the cardiac abnormality and maternal depression.
Infants subjected to cardiac surgery are at risk for impaired cognitive development,1 behavioural problems,2 attention problems, hyperactivity3 and emotional symptoms including anxiety.4 The mechanisms underlying these outcomes are complex, multifactorial5 ,6 and likely to involve biology and early experience as development reflects complex bidirectional transactional processes between the developing infant and the environment.7 The critical importance of a continuous, secure attachment relationship between mother and child, characterised by maternal warmth, sensitivity and responsivity,8 ,9 for cognitive and emotional development,8–12 optimal regulation of stress hormones13 and mental health14 is known for healthy children. High levels of stress associated with adverse psychosocial environments and insufficient buffering parental support are associated with lifelong negative outcomes (cognitive functioning, behavioural regulation, physical and mental health).15
Parents of critically ill children, regardless of diagnoses, are at increased risk of experiencing stress,16 depression, anxiety and traumatic stress responses6 ,17–19 which could impair their capacity to buffer the stress experienced by their infants. Parenting style and maternal mental health have been identified as important contributors to compromised child cognitive and behavioural outcomes in infants with congenital heart disease (CHD)6 although neurodevelopmental problems rather than maternal insensitivity has been associated with disorganised attachment in very preterm infants.20
To examine the impact of illness on the mother-child relationship, Goldberg compared 54 1-year-old children with CHD with 40 children with cystic fibrosis and 51 healthy controls using the Strange Situation Procedure (SSP) to measure attachment security.21 Both medical groups had fewer secure, more avoidant and more disorganised mother-child attachment relationships than the healthy group with the insecure attachment relationships associated with clinically significant behavioural problems at age 2–3 years.21 Impaired mother-infant interactions have also been found in 6-month-olds with CHD admitted to hospital for elective surgery22 and infants with CHD whose mothers have a highly anxious and avoidant attachment style have been found to have an increased risk of emotional problems at age 7 years.23 Researchers suggest that the SSP (valid for 1-year-olds) presupposes attachment already exists24 and that there is a need to study the unfolding of the infant-caregiver relationship.25 We hypothesised that CHD mother-infant dyads at risk of developing insecure attachment relationships might experience early relationship difficulties, and thus the aim of our study was to explore mothers’ subjective experience of their relationship with their infant soon after hospital discharge after surgery, using maternal self-report and complementary strategies of quantitative and qualitative data collection.26 We explored whether the mother-infant relationship quality was associated with medical variables or maternal depression.27
Patients and methods
Participants were infants, and consecutive admissions for cardiac surgery, at a tertiary centre before the age of 3 months, and their mothers. Eligible infants were identified by the treating medical team and recruited while inpatients. Written informed consent was obtained by the research assistant. Infants were excluded if mothers did not speak sufficient English to complete questionnaires and participate in the interview, did not live in Australia or if the infant was determined by the medical team to be too medically unstable for parents to be approached to participate.
Mothers completed the study questionnaire, which incorporated standardised measures, and were interviewed after the infant had been home from hospital for 4 weeks. Medical details (ie, timing of diagnosis) were obtained from the cardiac surgery database and infant's medical record. The cardiac specialists on the team rated the severity of illness (minor, moderate or severe) and prognosis (good, moderate or poor) for each participant.
The interviews were conducted by a graduate psychologist and the sequence and wording of questions was standardised. Some questions were outside the scope of this paper and are reported elsewhere.28 ,29 This paper reports on responses to two questions; “Do you think the surgery, illness and hospitalisation has impacted on your relationship with your baby?” followed by an open-ended question; “If yes, in what ways?”
As attachment refers to the emotional tie between the infant and caregiver as well as to the behavioural system30 we included a measure of mother's feelings of attachment to her infant; the Maternal Postnatal Attachment Scale (MPAS),31–34 a 19-item self-report questionnaire that assesses mother-infant attachment feelings in the postnatal period. Scores range from 19 (low attachment) to 95 (high attachment).31 The Edinburgh Post Natal Depression Scale (EPDS)35 is a 10-item self-report screening tool for postnatal depression. Responses relate to feelings over the previous 7 days. Scores range from 0 to 30 with higher scores indicating increased severity of depressive symptoms. A total score >12.5 indicates probable depression and a score ≥10 has been used in community samples to identify possible depression.36 ,37
Responses to interview questions were recorded in writing verbatim and transcribed into the study database. Two authors (BJ and NA) independently read each participant's response, identified themes using inductive thematic analysis and jointly developed a coding template incorporating the themes. Two coders were used to reduce the risk of a theme being overlooked due to bias.38 All the responses were then systematically coded by one researcher (BJ). Although codes were not identified in advance of the first reading of the data, it is inevitable that the identification of themes was informed by the clinical and theoretical perspectives (infant mental health and attachment theory) of the researchers.26
Questionnaire data were analysed with STATA V.11 software (Stata Corp, College Station, Texas, USA). Only complete data sets were used for each measure. Descriptive data are presented as means and CIs for normally distributed data and medians for skewed distributions. The relationship between infant or maternal characteristics and maternal attachment feelings (MPAS) and maternal depression (EPDS) was examined using OR for categorical data and t tests and Pearson's correlation for continuous data, unless skewed when the Wilcoxon rank sum test was used. Comparison with Australian norms for MPAS was done using one sample Student t test.
The thematic analysis of the interview data revealed the unexpected finding that almost a quarter of mothers had reported difficulties ‘bonding’ with their infant. We further explored this finding by generating a binary variable labelled ‘bonding difficulties’ and looking for possible associations between infant or mother characteristics and ‘bonding difficulties’ using χ2 and OR for categorical data and t test for continuous data.
Of the 198 infants who underwent cardiac surgery under 3 months of age between February 2005 and September 2006, 115 eligible families were approached (27 infants died prior to discharge, 17 were medically unstable, 9 families were uncontactable, 7 were unable to read English, 3 lived overseas and 20 could not be enrolled due to logistical difficulties). Seventeen of the 115 (15%) declined to participate and 1 infant died. Ninety-seven mothers agreed to participate. Seventy-eight (80%) returned completed questionnaires and 91 (94%) were interviewed. The majority of mothers (64/91, 70%), requested a telephone interview rather than an interview to coincide with attendance at outpatient review appointment. Characteristics of participating and non-participating infants and mothers were not compared due to restricted access to demographic data for non-participants under Australian privacy legislation.
In answer to the direct question about whether the illness, surgery and hospitalisation had impacted on their relationship with their baby, 56 (58%) mothers responded yes, 21 (22%) no, 10 (10%) initially and 4 were unsure. The thematic analysis of mothers’ responses to the second part of that question (“If yes, in what ways”) are presented below. As the aim of the qualitative data collection was to understand the findings and enrich the description of phenomena,26 themes are described below with illustrative quotes.
Eighty-one of the 91 mothers interviewed described their relationship with their infant including 8 who said there had been no impact and described the relationship in positive terms. Four main themes were identified in interview responses of mothers who said there had been an impact; the emotional tie (n=34, 37%), ‘bonding’ difficulties (n=22, 23%), anxiety and worry about the infant (n=17, 19%), and caregiving behaviours towards the infant (n=10, 11%). Themes were not mutually exclusive; 17 mothers encompassed two themes in their responses and 1's mother response encompassed three themes.
Within the emotional tie theme 15 mothers described a closer or stronger bond with their infant, another 10 described how precious the baby was and another 9 mothers described enhanced protective feelings (including 2 mothers who reported initially experiencing bonding difficulties). Some felt that the amount of uninterrupted time in hospital sitting with their infant contributed to a closer relationship.
Not in a negative way. I don't think so. If anything impacted in a positive way. We've all bonded in a positive way. It probably changes the way we deal with her.…. Closer to her. More thankful to have her
Responses were coded as belonging to the ‘bonding difficulties’ theme when mothers (n=13) used the word ‘bond’ to describe negative impacts on the relationship (eg, “felt bond not there”, “couldn't relax and bond with her”, “didn't feel we had a bond thing”) or used other language to describe similar strain in the relationship (n=9) for example, “I felt she wasn't my baby” “I was just a nurse that had a lot of shifts” “I didn't feel like her mother”. Mothers named separation at birth, intrusion of medical equipment, fragility of the infant, lack of opportunity to hold and do normal caregiving, sharing the care with nurses and mothers’ emotional responses of fear, anxiety and vigilance as contributing to bonding difficulties.
Lost confidence when it came to breastfeeding which is interesting considering he's my third. Made difficult to bond with him and even when home had trouble bonding with him. Really found it hard to bond.
Many (17/91, 19%), spoke about anxiety, worry and stress dominating the relationship.
Yes, more wary of everything, More aware and jumpy. If anything wrong—worried straight away. Overprotective, because been through so much.
Nine of the 91 (10%) mothers described more responsive caregiving behaviour.
I think we parent her very differently to how parented son. We don't let her cry for extended periods of time because think she could bust her gasket. With son did ‘controlled crying’ at 8 months but not prepared to do that with her…. We know we parent differently
The mean score on the MPAS (n=74) was 84.5 (SD 7.2, 95% CI 82.9 to 86.2) which is similar to Australian community norms for mothers of 4-month-old infants (mean 84.6, SD 7, p=0.47). Eleven mothers (15%) had scores lower than 1 SD below the community mean which we considered an indicator of ‘low attachment feelings’ likely to be of clinical interest as this score is lower than 85% of the community sample surveyed by the authors of the scale.
Mothers who described bonding difficulties had lower MPAS scores than those who did not (mean 80.6 (SD 10) vs mean 85.7 (SD 5.7), p=0.0047). Factors associated with bonding difficulties were prenatal diagnosis (OR 2.695% CI 0.89 to 8.9) and EPDS score (9.1 (5.3) vs 6.2 (3.9), p=0.01). Maternal depression was also associated with lower MPAS score (p=0.0001) (see table 3).
The aim of our study was to explore mothers’ subjective experience of their relationship with their infant soon after hospital discharge following cardiac surgery. As a whole group, mothers’ attachment feelings did not differ from community norms when measured using the MPAS. Thematic analysis of interviews provided a more nuanced picture. Most mothers reported a positive relationship with their infant. Almost half reported heightened attachment feelings and/or responsive caregiving, indicating that the attachment system was working well with medically ‘fragile’ infants eliciting increased protectiveness and care from their mother.
In contrast, almost a quarter of mothers interviewed indicated difficulty ‘bonding’ with their infant and this was associated with a prenatal diagnosis, high EPDS scores and low MPAS scores. These mother-infant dyads may be those at risk of developing the insecure attachment relationships documented by Goldberg.21 A fifth of our sample reported that anxiety, stress and fear dominate the relationship. A possible clinical explanation for the differing impacts of cardiac surgery on the mother- infant relationship is that the diagnosis of CHD is a crisis that can either mobilise adaptive coping resources or be associated with ongoing difficulties in the mother-infant relationship. The direction of the association between maternal depression and bonding difficulties is unknown. Prenatal diagnosis and length of stay were associated with bonding difficulties indicating that prenatal diagnosis may not always facilitate adjustment to the challenges of having an infant with CHD requiring surgery. Some mothers may (unconsciously or consciously) try to protect themselves from the fear of their infant dying by not forming a close bond with the infant. The limitations of our study include the lack of a control group, and lack of detailed data on preoperative, perioperative and postoperative medical factors. Caution needs to be exercised when generalising the findings due to lack of demographic data on non-participants, missing questionnaire data and the wide CIs for the associations between prenatal diagnosis and the EPDS cut-off scores and ‘bonding difficulties’. Ours was an exploratory descriptive study using thematic analysis of interviews. Future research on the early mother-infant relationship would benefit from directly asking mothers about the impact of the surgery on their developing bond with their infant, standardised observational assessments of mother-infant interaction in the neonatal period and longitudinal studies examining the relationship between these measures and MPAS scores and SSP classifications of attachment security at 1 year of age. The potential link between mother- infant relationship factors and the neurodevelopment, emotional and behavioural outcomes for infants with CHD could be investigated with an adequately powered study that combines a focus on the preoperative, perioperative and postoperative medical factors with a focus on the dynamics of the mother-infant relationship from prenatal diagnosis onwards. Our findings support the importance of early psychosocial care for infants with CHD advocated by McCusker et al.40 We recommend that prenatal and postnatal clinical care include screening, assessment and provision of early intervention for treatment for parent anxiety, depression and/or trauma stress responses and parent-infant relationship difficulties.
Psychosocial interventions to address the issues identified by mothers as contributing to bonding difficulties should be available from diagnosis onwards and encompass ‘universal’ psychoeducational interventions (eg, brochures, parent group counselling during admission) and targeted interventions for high-risk groups that focus on how to build and sustain relationships with infants in the context of hospitalisation. Treatment for parent mood or other mental health symptoms should also focus on the relationship with their baby. We recommend infant mental health treatment should be available for dyads where the mother-infant relationship is compromised. Greater focus on infant and family mental health at the time of surgery may have the potential to contribute to alleviating distress and preventing the cognitive, behavioural and mental health problems observed during childhood and adolescence.
Contributors BJ was involved in the design of the study, the management of the project, the supervision of the research assistant (data collection and analysis), the analysis of the data and the writing of the paper. BJ is responsible for the overall content as guarantor. CFR was involved in study recruitment, data collection, data analysis and writing the paper. NA was involved in data entry, qualitative data analysis and reviewing the paper. VA was involved in the design of the study, the supervision of the research assistant (data collection and analysis), and writing the paper. EN was involved in the design of the study, the supervision of the research assistant (data collection and analysis), and writing the paper. AC was involved in the design of the study, the recruitment of subjects, medical data collection, supervision of the research assistant, and reviewing the paper. SM was involved in the design of the study, the recruitment of subjects, medical data collection, supervision of the research assistant and writing the paper.
Funding We acknowledge grant support from Murdoch Childrens Research Institute, Australia and the Victorian Government's Operational Infrastructure Support Programme.
Competing interests None.
Ethics approval Royal Children's Hospital Melbourne Human Research Ethics Committee (HREC 21058).
Provenance and peer review Not commissioned; externally peer reviewed.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.