Women pregnant after previous perinatal loss: relationships between adult attachment, shame, and prenatal psychological outcomes

ABSTRACT Background Insecure adult attachment, shame, self-blame, and isolation following perinatal loss place bereaved women at risk of adverse psychological outcomes, which can impact child and family outcomes. To date, no research has considered how these variables continue to influence women’s psychological health in pregnancy subsequent to loss. Objective This study explored associations between prenatal psychological adjustment (less grief and distress) and adult attachment, shame, and social connectedness, in women pregnant after loss. Method Twenty-nine pregnant Australian women accessing a Pregnancy After Loss Clinic (PALC) completed measures of attachment styles, shame, self-blame, social connectedness, perinatal grief, and psychological distress. Results Four 2-step hierarchical multiple regression analyses revealed adult attachment (secure/avoidant/anxious; Step 1), shame, self-blame, and social connectedness (Step 2) explained 74% difficulty coping, 74% total grief, 65% despair, and 57% active grief. Avoidant attachment predicted more difficulty coping and higher levels of despair. Self-blame predicted more active grief, difficulty coping, and despair. Social connectedness predicted lower active grief, and significantly mediated relationships between perinatal grief and all three attachment patterns (secure/avoidant/anxious). Conclusions Although avoidant attachment and self-blame can heighten grief in pregnancy after loss, focusing on social connectedness may be a helpful way for prenatal clinicians to support pregnant women during their subsequent pregnancy – and in grief.

a stillbirth (death of baby after 20 weeks, or > 400 grams in birthweight), and two neonatal death (death of a baby within the first 28 days of life).Perinatal loss is traumatic (Armstrong, 2002), and the support of others is crucial for supporting women's psychological adjustment following loss (Astunda-Budak et al., 2015;Cacciatore et al., 2009).For a considerable number of women, her grief is disenfranchised (Doka, 2002); that is, the significance of the loss is not usually recognised, acknowledged or supported by health professionals, family, and society (Cacciatore et al., 2008;DeFrain et al., 1991;Duncan & Cacciatore, 2015).Women (and men) bereaved by perinatal loss frequently report social isolation and disconnection from others, as well as a sense of guilt, self-blame, failure, and shame (Burden et al., 2016).People with more insecurity in their relationships -avoidance and anxiety specifically -also report more shame, anxiety, depression, and loneliness (Akbag & Imamoglu, 2010;Gross & Hansen, 2000;Ruch, 1996;Russell et al., 2014;Wei et al., 2005).Even though more insecurity in adult relationships, shame, self-blame, and social isolation are associated with poorer psychological adjustment for bereaved mothers (Blackmore et al., 2011;Côté-Arsenault & Dombeck, 2001), it is not yet known how these key variables are associated with prenatal psychological adjustment after previous perinatal loss.Given more than 80% of women who experience perinatal loss will have a subsequent pregnancy within 18 months (Cuisinier et al., 1996), more research on the relationships between prenatal psychological adjustment and adult attachment, shame, and social connectedness, may help support women's prenatal psychological adjustment following loss.
Women pregnant after loss describe feeling like a failure because their body did not produce a healthy baby (Frøen et al., 2011) -and also question their ability to carry and birth a healthy living child in the subsequent pregnancy (Cacciatore, 2009;DeFrain, 1986).Although usually defined as guilt, feeling like a failure is more consistent with academic definitions of shame (Gilbert, 1998(Gilbert, , 2002)).In Gilbert's (1998Gilbert's ( , 2002) ) evolutionary framework, guilt and shame are defined as 'self-conscious emotions'; that is, an emotional experience linked to one's beliefs about others' perceptions of us.Shame is the negative evaluation of the self as a failure, undesirable, defective, worthless, and powerless (Gilbert, 1998(Gilbert, , 2002)).Unlike shame where one's self is evaluated as being bad, in guilt one's behaviour is the focus of negative evaluation.In a meta-analysis (Kim et al., 2011), shame was more strongly related to depression than guilt.Further research -utilising Gilbert's (1998Gilbert's ( , 2002) ) framework that defines shame from guilt -may help to deepen understanding of more helpful ways to enhance women's prenatal psychological adjustment following loss.
Social connectedness with others -a sense of belonging, closeness, and connection with others, the community and society at large (Lee & Robbins, 1995) -is crucial in supporting psychological adjusting following loss (Astunda-Budak et al., 2015;Cacciatore et al., 2009).Social connectedness develops best on a bedrock of attachment security (Bartholomew, 1990).Adults who have developed a secure (attachment) style of relating will be readily able to feel close, connected to, and supported by others, and believe that they are worthy of this care, connection, and support (Lee et al., 2001).Alternatively, people with an avoidant pattern of relating tend to view others as being uncaring or unavailable for connection, and people with an anxious attachment pattern view themselves as being unworthy of care and support, even when it is offered (Lee et al., 2001).Social connectedness has been shown to significantly, and fully, mediate the relationships between extraversion and well-being (Lee et al., 2008), and between self-esteem and social anxiety (Mahmooda et al., 2017).To date, no research has explored whether social connectedness mediates the relationship between adult attachment pattern and prenatal psychological adjustment following loss.In adults bereaved by the death of an older child, a secure pattern was related to improved psychological adjustment (less grief and distress), and an insecure (avoidant/anxious) pattern to poorer psychological adjustment (more grief and distress) (Barr, 2012).Exploring whether social connectedness could mediate the relationships between attachment pattern and perinatal grief may provide a possible intervening process to better support bereaved women pregnant after loss.

Present study
The aim of the present study is two-fold.First, to expand upon the limited body of research by exploring associations between prenatal psychological adjustment (less grief and distress) and adult attachment, shame, self-blame, and social connectedness, for women pregnant after loss.It was hypothesised that improved prenatal psychological adjustment would be associated with: less adult attachment insecurity, less shame and self-blame, and more social connectedness.Second, to explore social connectedness as a mediator of the relationships between attachment pattern and psychological adjustment.It was hypothesised that social connectedness would mediate the relationships between attachment pattern and psychological adjustment (less grief and distress).Examining these relationships may offer insights into further prenatal care pathways to support women's prenatal psychological adjustment which, in turn, may help her in her connectedness with others (including her relationship with her subsequent baby).

Participants
All procedures performed in this study were in accordance with ethical approval granted by the Mater Misericordiae Ltd Human Research Ethics Committee for the sub-study amendment (Contribution of shame to adult attachment in mothers who have experienced perinatal loss) to Project 20,308 (Evaluating a Pregnancy After Loss Service; Reference Number: HREC/13/MHS/7/AM10). Ethical approval was also granted by The University of Queensland Human Research Ethics Committee (Reference Number: 2018001332).Participants were 29 women in their subsequent pregnancy after a previous perinatal loss.This study used data already collected as part of a larger study evaluating the outcome of the Pregnancy After Loss Clinic (PALC) (Meredith et al., 2017).The PALC is a specialised midwifery-led clinic which provides individualised midwifery and obstetric care to couples who have experienced the death of a baby in their previous pregnancy.

Prenatal demographic details
The following maternal demographics were collected: loss variables (history of previous loss, time since loss, maternal age at loss, gestation of loss, history of multiple losses), previous mental health history, education level, and gestation at current pregnancy.

Prenatal adult attachment
The Attachment Style Questionnaire (ASQ; Feeney et al., 1994) is a 40-item self-report of adult attachment patterns.The measure has five subscales: Confidence (in self and others), Discomfort with Closeness, Need for Approval, Preoccupation with Relationships, and Relationships as Secondary.The subscales load onto three scales: Anxiety (Need for Approval and Preoccupation with Relationships); Avoidance (Discomfort with Closeness and Relationships as Secondary); and Secure (Confidence).The women were asked to rate, on a scale of 1 (Totally disagree) to 6 (Totally agree), how much they agreed with each of the statements.The ASQ has good psychometric properties (Feeney et al., 1994).Given the small number of participants in the present study, the three-factor model was used.Cronbach's α was adequate for the Secure scale at .67, and was excellent for the Avoidance (α =.84) and Anxiety (α=.83) scales.

Prenatal shame
The Shame Inventory (SI; Rizvi, 2010) is a 53-item self-report questionnaire that measures shame both globally and in specific situations.Only the first three items were used to measure global feelings of shame.The women were asked to rate, on a scale of 1 (Never/ None/No effect) to 5 (Always/Seldom/Extreme effect), how much they agreed with the statement.The complete scale shows good psychometric properties including internal consistency, test-retest reliability, and construct and predictive validity (Rizvi, 2010).Cronbach's α for the SI in the present study was .88,indicating excellent internal consistency.

Prenatal self-blame
The Self-Blame Questionnaire (SBQ; Bulman & Wortman, 1977) is a 5-item self-report questionnaire that measures an individual's perceived level of self-blame relating to a specific situation (perinatal loss).In the current study, the women were asked to rate, on a scale of 1 (Not at all) to 5 (Completely), how much they blamed themselves for their loss.Cronbach's α for the SBS in the present study was .90,indicating excellent consistency.

Prenatal social connectedness
The Social Connectedness Scale (SCS; Lee & Robbins, 1995) is a 29-item self-report questionnaire that measures social connectedness; that is, the experience of feeling close, and connected, to others.Questions relate to the perception (subjective awareness) of emotional closeness or distance between the self and others.The women were asked to rate, on a scale of 1 (Strongly disagree) to 6 (Totally agree), how much they agreed with each of the statements.The questionnaire has demonstrated good internal reliability (α=.91), excellent test-retest reliability (α=.96), and initial evidence of convergent and discriminant validity (Lee & Robbins, 1995).Cronbach's α in the present study for the SCS was .96indicating excellent internal consistency.

Prenatal perinatal grief
The Perinatal Grief Scale-33 (PGS-33; Potvin et al., 1989) was used to measure perinatal grief symptoms.The women were asked to rate, on a scale of 1 (Strongly agree) to 5 (Strongly disagree), how much they agreed with each of the statements.The PGS-33 shows good psychometric properties, including internal consistency (α=.94), test-retest reliability, and convergent validity.(Lasker & Toedter, 1991;Toedter et al., 2001) The scale includes 33 items split into three subscales that evaluate different dimensions of perinatal grief in relation to the lost baby: Active Grief, Difficulty Coping, and Despair.Cronbach's α's for the present study were excellent: .84(Active Grief), .90(Difficulty Coping), .90(Despair), and .94(Total).

Prenatal psychological adjustment
The Mental Health Inventory (MHI; Veit & Ware, 1983) is a 38-item questionnaire designed to measure general psychological distress and well-being in the general population.It contains six subscales: three measuring psychological well-being (Psychological Wellbeing Index: Life Satisfaction, Emotional Ties, and General Positive Affect), and three measuring psychological distress (Psychological Distress Index: Anxiety, Depression, and Locus of Control).Only the Psychological Distress Index was used in the present study given the small number of participants.The women were asked to rate, on a scale of 1 (Extremely happy/Always/All of the time) to 6 (Very dissatisfied/None/Never), how much they agreed with each of the statements.The MHI has good psychometric support for the distinct constructs, with good internal consistency (Veit & Ware, 1983).Cronbach's α for the Psychological Distress Index was .91 in the present study, indicating excellent internal consistency.

Statistical analyses
All analyses were conducted using the Statistical Package for Social Sciences (SPSS) version 26 with the significance level of .05.Missing values were computed using prorating to enable inclusion of all eligible cases in the analyses (Field, 2013).Pearson's correlations were used to investigate the relationships between the prenatal psychological adjustment (grief and distress) and adult attachment, shame, self-blame, and social connectedness.Based on previous research (Kirby et al., 2019;Kirby, 2016aKirby, , 2016b;;Mitchell et al., 2018), adult attachment was entered at Step 1 and shame, self-blame, and social connectedness at Step 2 for the regression analysis predicting prenatal psychological adjustment (grief and distress).The residuals for all regression models were checked for normality and homoscedasticity.Significant outliers and multi-collinearity were checked (Preacher & Hayes, 2004).Based on Cohen (1992), a sample size between 85 (correlation) and 97 (regression) was required to have sufficient power to detect a medium effect size for the correlation and linear regression (with six predictors) analyses.Given the small sample size, a bootstrap mediation analysis -PROCESS (Preacher & Hayes, 2004) -was used (N = 5,000 samples) to explore the influence of social connectedness on the relationships between attachment pattern (secure/avoidant/ anxious) and psychological adjustment (grief and distress).

Results
Demographic and sample characteristics are shown in Table 1.Most pregnant women were in their first trimester (72.41%) subsequent to perinatal loss.Nearly all had experienced a previous, single, loss (96.55%).Average gestational age at time of loss was 32 weeks, and mean time since loss was a little under 4 years.Almost women were in a relationship (96.3%), and just under half had other children (40.7%).The average maternal age was 32 years.Most of the pregnant women were Caucasian (81.5%).
Correlations are shown in Table 2.More avoidance in attachment was related to more difficulty coping and despair -and less social connectedness.More anxiety in attachment was correlated with more shame and perinatal grief (active grief, difficulty coping, despair) -and less social connectedness.More security in attachment was related to less shame, less self-blame, less grief -and more social connectedness.
Four 2-step hierarchical multiple regressions were conducted for active grief, difficulty coping, despair, and total grief.Given that psychological distress was not significantly correlated to the variables (attachment styles, shame, self-blame, social connectedness), it was excluded from regression models.As shown in Table 3, the model at Step 1 (attachment) was significant in predicting increased active grief, difficulty coping, despair, and total grief.The model at Step 2 (shame, self-blame, and social connectedness) explained significant further variance in active grief, difficulty coping, despair, and total grief.The total model explained 74% in difficulty coping, 74% in total grief, 65% in despair, and 57% in active grief.In these models, avoidant attachment predicted increased difficulty coping (p = .04)and despair (p = .04).Self-blame predicted increased active grief (p = .01),difficulty coping (p = .01),despair (p ≤ .001),and total grief (p ≤ .001).Higher social connectedness predicted lower active grief (p = .02).As shown in Figures 1 to 3, social connectedness fully mediated the relationships between secure attachment and total grief (p ≤ .001),avoidant attachment and total grief (p ≤ .001),and anxious attachment and total grief (p ≤ .001).Given the small sample size, only total grief was used as an outcome variable.Neither shame nor self-blame were significant mediators of the relationships between adult attachment and grief.

Discussion
The aim of the present study was two-fold.The first aim was to explore associations between prenatal psychological adjustment (less grief and distress) and adult attachment, shame, self-blame, and social connectedness, for women pregnant after loss.It was hypothesised that improved prenatal psychological adjustment would be associated with: less adult attachment insecurity, less shame and self-blame, and more social connectedness.Overall, most findings supported the hypothesis.Pregnant women with higher levels of avoidant attachment reported more difficulty coping and higher levels of despair, whereas pregnant women with higher anxious attachment reported more shame and perinatal grief (active grief, difficulty coping, despair).Both reported lower levels of social connectedness.The findings of this study are consistent with previous research which shows shame and self-blame are associated with heightened perinatal grief (Christiansen et al., 2013;Shevlin et al., 2014).The current study extends this research by showing that an avoidant attachment pattern and self-blame are associated with higher levels of grief in the subsequent pregnancy after loss.The findings that avoidant attachment was associated with higher levels of shame and self-blame is consistent with previous research for parents bereaved by the death of their child (Huh et al., 2017).The second aim was to explore whether social connectedness could mediate the relationships between attachment and psychological adjustment.The finding that social connectedness fully mediated the relationships between attachment pattern and grief was a unique finding of this study.The results of the present study broaden our understanding by highlighting that an avoidant attachment pattern, self-blame, and social connectedness can continue to influence prenatal psychological adjustment following perinatal loss (Cacciatore et al., 2009), and social connectedness may provide a unique avenue for enhancing prenatal psychological adjustment.There were a few findings that were unexpected.First, the results of self-blame as a significant, and unique, predictor of perinatal grief is consistent with previous research (Cacciatore et al., 2013;Duncan & Cacciatore, 2015); however, the finding that shame was not a predictor is inconsistent with results from previous studies.This may be due to the supportive relationship that may have developed between the pregnant woman and the midwives in this PAL Clinic.It was not uncommon for the pregnant woman to have previously met the PALC midwife at the time of the baby's death.Alternatively, the measure of self-blame and the difficulty coping subscale of the perinatal grief scale (e.g.I blame myself for the baby's death, I feel guilty when I think about the baby, the best part of me died with the baby) may have been more sensitive to shame experienced by women in perinatal grief.Second, the finding that psychological distress was not significantly correlated with shame, self-blame, attachment styles, or perinatal grief is inconsistent with expectations and with previous research, particularly for the relationship between shame and depression (Irons et al., 2006).It is likely that the current study may have lacked statistical power to detect a significant effect given the small sample size (29 versus 85 required).Alternatively, a floor effect may have impacted results.Although normative values for psychological distress (from the MHI) have not been published for Australian populations, the mean value of distress in the current study was substantially lower than the mean value for the normative US population (Veit & Ware, 1983).Only one of the pregnant women in this study had a clinically elevated score on the psychological distress scale.It is also possible that the perinatal grief scale may have tapped into psychological distress specific to perinatal grief, compared with a general measure of distress.In previous research, a little more than half the variability in depression was explained by perinatal grief, specifically the difficulty coping subscale (Potvin et al., 1989;Toedter et al., 2001).Whether the perinatal grief scale is identifying women at risk of adverse grief outcomes is a potential area of future research.

Limitations
Results of this study should be considered in light of several limitations.First, the study sample size was small and, like previous research in this field, the sample consisted mostly of married, educated, Caucasian women with high levels of education.The women in this study also had previous losses of stillborn and neonatal death.Therefore, the results may not be applicable to women from different cultures/SES or who experienced a miscarriage, multiple losses, or ended a pregnancy early due to foetal abnormality.Future research would benefit from engaging women with different loss experiences, and a variety of socioeconomic, educational, and ethnic backgrounds.Second, the sample size in the current study was too low to detect a medium effect size.Despite the significant, and clinically promising, findings, future research with a larger sample, comparing mothers with and without perinatal loss, is needed to increase confidence in these findings.

Clinical relevance
In terms of prenatal care, these findings identify areas of potential relevance.First, clinicians who support women during their pregnancy subsequent to perinatal loss should be aware of the potential for perinatal grief even in the absence of psychological distress.Pregnant women should be monitored throughout the subsequent pregnancy in relation to their grief: how they are managing their grief, if they are experiencing selfblame, and how they are viewing their relationships with others in their lives.Pregnant women with an avoidant or anxious style of relating may need support in their relationships with others, including their developing maternal-foetal attachment, during their pregnancy and beyond.Pregnant women with an avoidant style of relating may need further support in being able to approach and turn towards others when distressed, rather than minimising or denying the extent of their emotional needs and grief.Increased levels of distancing of themselves from others, or being critical of others (and themselves), may indicate further intervention is needed.Pregnant women with an anxious style may be more attuned to how other people in their life are responding to the subsequent pregnancy (husband/partner, mother, other children).If they are focusing more on how other people in their lives are coping, or not coping, to the exclusion of their own needs, this may be indicative of a cause for concern for women with an anxious style.Interventions that focus on cultivating mindfulness and self-compassion may be helpful in supporting pregnant women in enhancing their relationships with others; specifically, their family and support networks, their relationship with the baby that died, and their subsequent pregnancy (Gammer et al., 2020;Lennard et al., 2021;Mitchell et al., 2018).

Conclusions
Results of this study support the larger body of research showing that a history of previous perinatal loss may be related to prenatal psychological adjustment in the subsequent pregnancy.Pregnant women with a history of insecure attachment (avoidant/anxious) were shown to have higher levels of perinatal grief, and lower levels of social connectedness.Pregnant women who blamed themselves for their loss (self-blame) was shown to significantly predict increased perinatal grief.The study highlighted the importance of social connectedness in mediating the relationships between attachment patterns (secure, anxious, avoidant) and grief.These findings indicate that incorporating a focus on increasing a sense of social connectedness with others may be beneficial for clinicians supporting pregnant women in managing their grief and pregnancy.Although further research is needed to explore the applicability of these findings to a larger population of women pregnant after loss, these findings provide potential areas for practical intervention for clinicians supporting women pregnant after loss.

Table 2 .
Descriptive statistics and correlation matrix for the main study variables, N = 29.