{PSOCIAL}

Bowlby’s (1973, 1982) attachment theory provides a coherent and empirically-validated conceptual framework for understanding both the quality of parent-child relationship and individual differences in grief responses to the death of loved ones. In Bowlby’s (1973) view, early interactions with responsive and supportive caregivers, usually the parents, in times of need contribute to the formation of an inner sense of security (i.e., confidence that one is socially valued and that others will be available and helpful when needed) and positive beliefs about the self and others. However, when caregivers are rejecting, unreliable, or focused on their own needs and distress, the sense of security is destabilized and attachment-related worries and anxieties tend to emerge in the form of insecure attachment (Bowlby, 1973).

To date, attachment orientations are conceptualized in terms of a two-dimensional space defined by two roughly orthogonal dimensions, attachment anxiety and attachment-related avoidance (e.g., Brennan et al., 1998). Attachment anxiety reflects worries that others will not be available when needed; doubts about one’s lovability and worth; and energetic, angry, and sometimes coercive, searching for support, love, and care (what Cassidy & Kobak, 1988, called hyperactivating strategies). Avoidant attachment reflects distrust in others’ intentions, compulsive self-reliance, dismissal of needs for relatedness and intimacy, and suppression of attachment-related thoughts and feelings (what Cassidy & Kobak, 1988, called deactivating strategies). People who score low on both dimensions have a prominent sense of attachment security. There are reliable and valid self-report scales to measure these two dimensions (e.g., the Experiences in Close Relationships scale, Brennan et al., 1998) .

There is evidence that attachment orientations are relevant for explaining individual differences in grief reactions. Secure attachment allows a person to work through the loss experience and return to normal functioning (Fraley & Shaver, 2016). Secure people can recall and think about the lost partner without extreme difficulty and can discuss the loss coherently without being overwhelmed by pain and distress (Hesse, 2016). In contrast, attachment-anxious people’s overdependence on the lost person’s attention and care might heighten distress, complicate the grief process, and delay return to normal functioning (Mikulincer & Shaver, 2013). Indeed, several studies have shown that attachment anxiety contributes to heightened and more prolonged distress following the death of a loved person (Field & Sundin, 2001; Jerga et al., 2011). For example, Fraley and Bonanno (2004) found that attachment anxiety assessed 4 months after the loss of a spouse predicted higher levels of distress, trauma-related symptoms, and alcohol consumption 18 months following the loss.

Adult attachment studies found no association between avoidant attachment and heightened distress following the death of a relationship partner (Field & Sundin, 2001; Fraley & Bonanno, 2004), but Wayment and Vierthaler (2002) found that avoidance was associated with increased somatic symptoms during the grief process. Interestingly, Jerga et al. (2011) found that the lack of avoidance-distress association was explained by the extent to which avoidant people felt close with the deceased partner. Avoidant people seem to maintain a cold and distant relationship with the deceased, which in turn leaves them with less to grieve about.

Although all the studies conducted on the attachment-grief link have focused on the death of a spouse, we hypothesize that attachment anxiety would also heighten distress of mid-life adults coping with the death of a parent. This is a novel hypothesis as no published study to date has examined the role of attachment insecurities in the context of parental death during mid-life.

As mentioned earlier, a person’s attachment orientations seemed to depend on parents’ sensitivity and responsiveness to his or her needs for protection and support in times of need. Research has shown that parents’ trauma-related history can lead to major difficulties in properly fulfilling these parental duties (De Haene et al., 2010; Schwerdtfeger & Goff, 2007). In some cases, due to their feelings of powerlessness, as they failed to protect themselves, traumatized parents cannot see themselves as a secure base for their child (Almqvist & Broberg, 2003). In addition, traumatic experiences can heighten parents’ psychological problems and then impair their ability to function as a sensitive and responsive attachment figure (Tambelli et al., 2015).

Much have been also written about Holocaust survivors’ complicated relationship with their offspring (Wiseman & Barber, 2008). Persecuted by their horrible war memories, many of Holocaust survivors saw the world as dangerous and tended to be anxious and overprotective to their offspring (Bar-On et al., 1998; Brom et al., 2001). In some cases, the surviving parents tend to be overwhelmed by their traumatic memories and then they do not have the necessary resources to function as a sensitive and responsive figure (Brom et al., 2001; Shrira et al., 2019). In other cases, researchers have documented a role reversal between parents and children, by which the HSO attempts to protect the traumatized parent (Shrira et al., 2019).

Now days, more than seventy years after the end of WWII, most of Holocaust survivors who are still alive are in their 80s-90s under the care of their offspring. Caring for an aging parent may be particularly demanding and challenging for HSO (Shmotkin et al., 2011; Shrira et al., 2019). Research indicates that HSO describe their parents as particularly vulnerable to aging-related changes, such as illness, frailty, loneliness, and dependency (Shmotkin et al., 2011). Simultaneously, old-age Holocaust survivors tend to report more frequent reactivation of Holocaust-related traumatic memories and unresolved losses (Shmotkin et al., 2011). Hence, HSO are not only burdened by the hard work involved in caring for an aging parent, but also by an urgent need to protect them from further pain, which can lead to guilt feelings for parents’ suffering and suppression of anger towards them (Wiseman & Barber, 2008).

A still more challenging event for HSO is the death of their old-age survivor parent. Although caring for an aging parent might be quite arduous, HSO can still maintain a vital relationship with the traumatized parent. In contrast, HSO who have lost their parent, although being “liberated” from the daily hassles and burden involved in caring for an aging parent, are confronted with the “final separation” from the traumatized parent and their irrevocable failure to protect him or her from death. Moreover, parental death can be particularly distressing for HSO, who care for their parents with great devotion, partly due to their ongoing need to compensate for parents’ past losses and partly due to their guilt feelings (Kellermann, 2009). In addition, Holocaust survivors have been found to be ambivalent over emotional expression (Ben-Zur & Zimmerman, 2005) and to convey a message that in order to survive one needs to keep emotions bottled up and exhibit a tough façade (Wiseman & Barber, 2008). This ’conspiracy of silence’ seemed to refrain HSO from asking their parents about Holocaust experiences, thus building a "double wall of silence" (Bar-On et al., 1998). As a result, a painful void can emerge after the death of the survivor parent and the untold haunting memories might heighten difficulties in calmly accepting parent’s death, completing a successful grief process, and restoring emotional equanimity (Brom et al., 2001).

Interviews with Holocaust survivors also revealed that while expecting their children to enjoy life, they also passed on a sense of obligation to the dead, which might pose further difficulties to HSO’s grief (Shrira et al., 2019). On this basis, we hypothesize that HSO, as compared to offspring of parents who did not experience the Holocaust atrocities, would react with heightened distress to the death of a parent. We also hypothesize that the detrimental effects of attachment anxiety would be more pronounced among HSO who might find harder to accept parent’s death

The sample included 146 Israeli adults (52 men and 94 women), ranging in age from 40 to 71 years (M = 55.2, SD = 7.81), who volunteered to participate in the study without receiving any monetary compensation. This specific age range was chosen due to two main reasons. First, many people in this age range are dealing with the loss of their parents (Hooyman & Kramer, 2006). Second, holocaust survivors went through World War II (WWII) during their adolescence or young adulthood and therefore could remember experiences during this period and share them with their offspring. All participants were Israeli-born Jews and their parents were born in Europe or Israel. Most participants were married (120, 82%), secular (120, 82%), with academic education (135, 92.5%) and average or above income (126, 86.3%).

We sampled two groups of participants: (a) participants who were grieving for one of their parents who died in the last three months before the study (grieving group, n = 61) and (b) participants who did not lose a parent in the last three months (non-grieving group, n = 85). In the grieving group, the other parent was alive at the time of the study or died over a year before. In the non-grieving group, both parents were alive at the time of the study or one parent died over a year before. In each of these groups, we attempted to equally represent HSO and non-HSO participants. Holocaust survivor offspring (n = 32 in the grieving group and n = 40 in the non-grieving group) were defined as participants whose at least one parent (the dead parent among grievers) was at concentration camps, ghettos, or was partisan during WWII, thereby reducing the heterogeneity of the Holocaust experiences. Non-HSO (n = 29 in the grieving group and n = 45 in the non-grieving group) were defined as participants whose parents immigrated to Israel from Europe before 1939 and their immediate family had no Holocaust background.

There was a significant difference in age between the two groups, t (1, 14 4 ) = -3.22, p = .002. Participants in the grieving group were older (M = 57.57, SD = 6.41) than those in the non-grieving group (M = 53.48, SD = 8.30). No significant age difference was found between HSO and non-HSO, t (1, 144) = -.82, p = .41. In order to control for this unexpected group difference, we introduced participants’ age as a covariate in the analyses examining the study’s predictions. No other significant differences were found between study groups in gender, marital status, education, religiosity, and income.

In the grieving group, 50% of the participants reported that their father died in the last three months before the study and 50% reported that their mother died during this period. In 36.1% of grieving participants, the other parent was alive at the time of the study. In the non-grieving groups, 45.9 % of the participants reported that both parents were alive at the time of the study.

Twenty-five percent of HSO reported that only the father was a Holocaust survivor, 18.05% reported that only their mother was a survivor, and 56.95% reported that both parents survived the Holocaust. 50% of them were at concentration camps, 43% were at ghettos, and 7% were partisans during WWII. No significant differences in these variables were found between the grieving HSO and non-grieving HSO groups, χ2s < 6.01, p > .06.

After receiving ethic approval from the university, participants were recruited using direct person-to-person solicitation in social clubs and community centers. In addition, we posted on-line advertisements and called for participation through social networks. We sought first for HSO participants and conducted a phone interview about their parents’ Holocaust background, whether their parents are alive, and whether one parent deceased during the last three months. We then asked participants who filled the inclusion criteria their permission to participate in the study. After collecting data in the HSO groups, we sought for controls who matched HSO in age and gender and conducted a similar phone interview to check whether they filled the inclusion criteria. From these phone interviews, 182 participants fitted the inclusion criteria, but only 146 agreed to participate in the study (rejection rate of 19.79%).

In this section, we report findings from two hierarchical regressions examining main effects and interactions of parental death, being a HSO, and attachment anxiety (unexplained by attachment avoidance) on psychological well-being and psychological distress while controlling for age as a covariate (see relevant statistics in Table 1). The regression model for well-being was significant and explained 30.03% of the variance (see Table 1). In Step 1, the main effects for parental death and attachment anxiety were significant. Fitting the study’s prediction, participants who were grieving for a parent’s death reported lower levels of well-being than non-grieving participants. In addition, participants scoring higher on attachment anxiety also reported lower levels of well-being (see Table 1). In Step 2, we found significant interactions for parental death x HSO and parental death x attachment anxiety (see Table 1). However, these two-way interactions were qualified by a significant three-way interaction in Step 3 (see Table 1).

In testing the study’ prediction, Simple Slope Effects tests revealed that grieving participants reported lower well-being than non-grievers only if they were HSO and scored relatively high on attachment anxiety (+1 SD), b = -.36, SE = .13, t = -2.75, p = .006. However, among non-HSO who scored relatively high on attachment anxiety (+1 SD), grieving participants reported higher well-being than non-grievers, b = .34, SE = .14, t = 2.48, p = .014. Among HSO or non-HSO who scored relatively low on attachment anxiety, grieving participants did not significantly differ from non-grievers in well-being, bs < -.24, all ps > .070. As can be seen in Figure 1, the lowest level of well-being was reported by HSO who were grieving a parent’s death and scored high on attachment anxiety. But unexpectedly, non-HSO who scored high on attachment anxiety seemed to benefit (heightened well-being) from the death of a parent.

The regression model for psychological distress was also statistically significant and explained 37.60% of the variance (see Table 1). In Step 1, the main effects for parental death and attachment anxiety were significant. Fitting the study’s prediction, participants who were grieving for a parent’s death reported higher levels of distress than non-grievers. In addition, participants scoring higher on attachment anxiety also reported higher levels of distress (see Table 1). In Step 2, we found a significant interaction for HSO x anxiety, which was qualified by a significant three-way interaction in Step 3 (see Table 1).

Simple Slope Effects tests revealed that that grieving participants reported higher distress than non-grievers only if they were HSO and scored relatively high on attachment anxiety (+1 SD), b = .25, SE = .11, t = 2.33, p = .021. Among non-HSO who scored relatively high on attachment anxiety (+1 SD) or among HSO or non-HSO who scored relatively low on attachment anxiety, grieving participants did not significantly differ from non-grievers in distress, bs < .14, all ps > .208. As can be seen in Figure 2, the highest level of distress was reported by HSO who were grieving a parent’s death and scored high on attachment anxiety.

Our study has several limitations. The main limitation was the difficulty to recruit participants to take part in a study dealing with parental death right after the loss, when the hardship and the pain are still so fresh. In addition, since the population of Holocaust survivors is decreasing dramatically each passing day, it was even more challenging to recruit HSO for both non-grieving and grieving groups. These difficulties have resulted in a relatively small sample, which restricted the study’s statistical power and the generalizability of findings. Additionally, our findings were based on self-report measures, which can be biased by social desirability, and a cross-sectional design that prevented us from making strong inferences about direction of causality. Further studies should include behavioral observations or less explicit measures of attachment, well-being and distress and rely on prospective longitudinal designs that can follow the contribution of parental death and attachment orientations to mental health over time.

In the current study we focused on HSO’s attachment orientation as a proxy of the relationship satisfaction they had with the traumatized parent. Future studies can benefit from examining other variables that can highlight the quality of such a relationship, such as parent’s attachment orientations or specific aspects of the relationship (e.g., intimacy, self-disclosure, investment). Also, future studies using larger samples can consider if both parents are dead or one of them is still alive. In addition, since elderly parental death has hardly been investigated until now, it is important for future research to assess how this event is perceived by their grieving children. This assessment is important due to previous findings showing an association between how much the event is evaluated as traumatic and the emergence of emotional problems mainly among attachment-avoidant people (e.g., Ogle et al., 2016).

The current study has several theoretical and clinical implications. Theoretically, it reinforces the understanding of grief reaction after parental death among mid-life adults. The addition of dead parent’s background and offspring’s attachment orientations takes this knowledge one step further by highlighting how factors that related to parent-child relationship contribute to the emergence of emotional and relational problems following this loss.

Clinically, the current findings reinforce the importance of developing specific intervention programs for mid-life people who are dealing with elderly parental death. As mentioned earlier, little is known about the difficulties this population is dealing with. It is important to consider their needs for special adjustments when they turn to treatment. However, in most cases therapists and therapeutic centers are not qualified or knowledgeable about those necessary accommodations (Laidlaw & Pachana, 2009).