Abstract
The death of a loved one can be heartbreaking for those left behind, and indeed, bereavement is associated not only with adverse health effects but also a higher risk of dying oneself. Not surprisingly, its consequences have been the subject of much psychological enquiry, with a major interest in shedding light on how one adapts, who is most at risk, and why. Often the focus is on the bereaved individual, yet people do not typically grieve in isolation; most do so with family members who have likewise experienced the loss. Family dynamics affect personal grief and vice versa. What is more, family concerns, such as reduced finances, legal consequences, and changed family relationships, have to be dealt with. While the latter stressful aspects have been investigated, there is still a huge gap between the individual and family approaches. To move them closer together, we propose a family-level extension of our Dual Process Model, showing how the whole may actually be more—and more accurate—than the sum of the two parts.
Two decades ago, Gilbert (1996) expressed an insightful opinion: “In order to truly understand the nature of grief in families, it is necessary to recognize that both individual and relational factors are operating and that these must be considered simultaneously” (p. 271). Unfortunately, her words seem to have fallen on deaf scientific ears; none of us have seriously taken up the challenge to actually integrate individual and family perspectives. Rather, researchers have come to understand that both intra- and interpersonal factors affect grief and grieving and have gratefully applauded complementary efforts in each area while continuing to study intra- and interpersonal phenomena largely independently. That a body of research in each area is developing is beyond question (see M. Stroebe, Schut, & Finkenauer’s, 2013, review). To illustrate: Scientists have expended much effort in empirical research to investigate coping at the individual level (e.g., looking at strategies of confronting vs. avoiding reminders of the loss) separately from examining ways of grieving with others (e.g., talking or not talking about it), occasionally also looking rather independently at interactive coping processes (e.g., the influence of one person’s grieving on that of another and how this affects adaptation). Similarly, there is a clear divide between individual-level theories of coping with bereavement (e.g., application of cognitive stress theory; Folkman, 2001) and family-level approaches (e.g., Nadeau, 1998; Walsh & McGoldrick, 2013), while recently there has also been growing recognition of the need to understand the impact of families on individuals and vice versa (cf. Dyregrov & Dyregrov, 2008; Kissane & Parnes, 2014; Rosenblatt, 2013; M. Stroebe, Schut, & Finkenauer, 2013).
Why then, one might ask, is it necessary to go a step further and actually, systematically, incorporate research and theorizing across these individual and family/interpersonal domains? Is what Gilbert said actually true—that “these must be considered simultaneously”—or are the parallel lines of investigation indeed sufficient? In our view, Gilbert was right, because (as we illustrate next) there is striking evidence that integration of individual- and family-level perspectives could deepen scientific understanding and increase the power and accuracy of prediction regarding the adjustment of bereaved persons. Continued efforts in this direction are crucial, given the mental and physical health consequences of this major life event and the need—particularly for the provision of effective health care—to better establish who among the bereaved is most vulnerable, and why (M. Stroebe, Schut, & Stroebe, 2007).
Recent reviews have highlighted concrete ways in which individual and family dynamics “fuel” each other during bereavement (Albuquerque, Pereira, & Narciso, 2015; M. Stroebe, Schut, & Finkenauer, 2013). That an integrative perspective may even have incremental value was demonstrated in the results of a study of our own (we draw on this study given the absence—to our knowledge—of other examples). This was a longitudinal investigation of couples coping with the death of their child (e.g., M. Stroebe, Finkenauer, et al., 2013; Wijngaards et al., 2008). Two hundred nineteen bereaved mothers and fathers were individually interviewed, and each of them provided written responses to questionnaires 6, 13, and 20 months after their child had died. Of particular relevance here, the impact of partner-oriented self-regulation (POSR) was assessed. POSR refers to a dynamic, interpersonal phenomenon characterized by the avoidance of talking about the loss of the child and remaining strong in the partner’s presence in order to protect the partner. Paradoxically, we found that holding in one’s own grief in order to protect one’s partner from pain was actually associated with greater grief for both the partner and the self later on. So, although parents tried to protect their partners through holding in their emotions, this attempt actually had the opposite effect, harming not only their partner but themselves as well. Shakespeare’s lines from Macbeth seem to resonate: “Give sorrow words. The grief that does not speak whispers the over-wrought heart and bids it break” (1623, 4.3.12–13).
Importantly in the current context, either an individual-level or couple-level investigation alone would have provided limited knowledge and could have led to faulty, even wrong, conclusions. Certain patterns found in the results are indicative. To start with, additional variance in grieving could be explained through this integrative approach: Predictors of grief over time occurred at the two levels; both self-reported and partner-reported POSR predicted grief after a longer duration of bereavement. Furthermore, these patterns pertained to both bereaved fathers and mothers, even though husbands held in their grief more for the sake of their wives than vice versa (consistent with findings from individual-level studies showing that fathers disclose grief less than mothers). The gender difference was shown, then, to be limited to a difference in averages between fathers and mothers. Had the focus been restricted to individuals, we might have concluded that when parents’ POSR impedes their own grief, they pay a personal price for this protective restraint toward their partner. By including both, and looking further at the interactive as well as intrapersonal processes, we showed that despite the gender difference in expressing grief, POSR comes at a cost for both parents. So, the whole picture that emerges is more than the sum of the parts—with the incremental value that by including these family matters, we begin to understand a little better who is at risk and why. We would never have discovered that the protection of the partner, meant to be helpful, has this paradoxical effect had we not included both individual and family levels of analysis.
How can we promote such integrative research? We are firm believers that theory should guide research: What we need is an integrative coping model. So far, though, coping models have focused on the individual level, failing to systematically cover influential factors in the family domain. To give them their due, acknowledgment of interpersonal influences is usually made in individual-level approaches, but this is a far cry from integration. Next, we outline how we went about developing a more structured individual–family coping perspective.
Development of the Dual Process Model and Integration of Family Tasks
The integrative model that we now propose extends our own Dual Process Model of Coping with Bereavement (DPM; M. Stroebe & Schut, 1999, 2010). The DPM owes much to earlier perspectives. So, let us step back through history for a moment to give the broader picture: A sequence of fascinating scientific contributions took place over the course of the 20th century surrounding the notion that one “has to do one’s grief work,” which were based on different principles but commonly inspired by the goal of understanding how people cope with bereavement. There were three major landmarks. First, the idea that one has to deal with the loss—to go through a grieving process—is typically traced back to “Morning and Melancholia,” the classic article by Freud (1917/1957). This process was said to enable the bereaved person to withdraw energy from the deceased and invest it in other relationships. Second, “grief work” featured prominently in Bowlby’s (e.g., 1980) attachment theory, reflecting the effort to regain proximity upon separation from the loved one—which, in the event of death rather than temporary separation, cannot be achieved. Bowlby postulated four (flexible, overlapping) “working through” phases, whereby the bereaved person comes to terms with irrevocable loss: shock, yearning and protest, despair, and recovery. Third, Worden’s (1982, 2009) so-called task model more explicitly emphasized a fundamental principle: Active, effortful working through of grief is necessary; one has to grapple with the reality and undertake certain tasks to come to terms with the death and ultimately adapt to life without this person. The postulated tasks were fourfold, too: accepting the reality of loss, experiencing the pain of grief, adjusting to life without the deceased, and relocating him or her and moving on.
The DPM was designed to overcome certain limitations of these previous grief-work models (see M. Stroebe & Schut, 1999, 2010). Importantly in the current context, such models had focused on adjusting to the loss of the loved person per se, not on additional upheavals that occur when someone dies: The surrounding world needs reorganizing, too, and this is upsetting and stressful. 1 These secondary stressors were built into the DPM, making it a “dual”-process model: attending to loss, on the one hand, and to change, on the other. There are tasks to work through in both these domains. The structural components of the DPM incorporate these tasks. They are depicted in Figure 1, alongside the new family-level extensions, combining these elements into the revised model (the Dual Process Model–Revised; DPM-R).
The Dual Process Model–Revised (DPM-R), which integrates loss-oriented and restoration-oriented tasks at both the individual and the family level. The individual-level loss-oriented tasks are based on Worden (1991). Worden made subsequent revisions, but this version was available at the time of the DPM’s development, providing the source for the DPM task extension. The third and fourth family-level restoration tasks are somewhat overlapping (and also overlap with the loss-oriented third and fourth tasks, as discussed next), but they occur—at least to some extent—in temporal sequence. To the extent that family-level restoration tasks are directly compensating for the loss of the family member, they may actually be considered loss oriented, but in so far as they lead to a review of family habits and general revision of traditions or ways of going about things in the family context, they can be considered restoration oriented: as ways to go forward. Difficulties may arise not only in relationship to the tasks themselves but from family members’ differences—or conversely, similarities—in oscillation (e.g., needing to confront/distract from grieving at different times, leading to incompatibility, vs. grieving intensely at the same time, leading to intensification of suffering).
To elaborate, the original DPM defined the two categories of stressors outlined above; both types have to be dealt (coped) with. Loss orientation refers to the bereaved person’s concentration on, appraisal of, and processing of stressful aspects of the loss experience itself. Worden’s (1991) four DPM-compatible tasks are included in the loss-oriented, individual-level tasks column. Restoration orientation refers to the DPM’s additional focus on other stressors, aspects that have come about as secondary consequences of bereavement. Figure 1 lists the individual-level restoration-oriented tasks, too, paralleling the loss-oriented ones. These have to do with efforts to reorient oneself in a changed world, relating to the disruption of one’s life due to the death. The business of restoration can also be associated with negative outcomes such as distress and anxiety. In addition to specifying such stressors, the DPM describes coping. This refers to processes, strategies, or styles of managing (reducing, mastering, tolerating) the (perceived) situation in which bereavement places the individual (cf. Folkman, 2001). Figure 1 depicts a dynamic, regulatory coping process: oscillation. At times the bereaved will confront aspects of loss, at other times avoid them, and the same applies to restoration stressors. Sometimes, too, there will be “time out,” when the person is not working through the loss. Oscillation is deemed necessary for adaptation.
Moving toward integration, we reasoned that if such tasks are necessary for successful adaptation at the individual level, a valid strategy would be to explore their viability at the family level. So, the extension of the DPM to the family level in the DPM-R parallels the individual-level structure. Family-level stressors refer to jointly experienced family-level matters that relate to adjustment of the family as a whole (and consequently its individual members). Family-level coping refers to working through individual- and family-level stressors as a family. 2 Families coping together can aid (or hinder) adjustment at both the individual and family level. Similarly to those at the intrapersonal level, loss-oriented family tasks have to do with coming to terms with the death of the loved person him- or herself (e.g., informal gatherings of family members can provide ways of undertaking loss-oriented tasks together; cf. Walter, 1996). Restoration-oriented family-level tasks have to do with coming to terms with subjective and objective changes within the surrounding world, those that came about as a result of the death (e.g., dealing together with reduced family income; see Corden & Hirst, 2013). Finally, following the DPM-R, it should be noted that consequences (outcomes) of the coping process can be experienced at both the individual and family levels (e.g., both personal health or loneliness and family conflict or divorce; cf. Albuquerque et al., 2015).
What evidence is there to support our integrative DPM-R? “Translating” the POSR study findings described earlier into task terms, one could argue that it illustrates both the second loss-oriented individual- and family-level tasks of experiencing and sharing the pain of grief. Practicing what we preach, we conducted further analyses of this data set to explore other types of related intra- and interpersonal dynamics (Buyukcan-Tetik, Finkenauer, Schut, Stroebe, & Stroebe, submitted manuscript). This time, the focus was on family adjustment—specifically, on changes in the couple’s relationship and the partners’ perceptions of both their own and the other’s level of grief (thus, family-level coping with the third restoration task). Importantly, actual (dis)similarity in level of grief was controlled for (i.e., by comparing the partners’ scores on the grief measure). Even with this comparative intrapersonal control, interpersonal dynamics played a part: Parents who perceived their grief to be dissimilar (less or more than their partner) had lower relationship satisfaction than those who perceived that they and their partner experienced similar levels of grief. These negative effects of perceived dissimilarity increased over time, with the decrease in relationship satisfaction being stronger among those who perceived that they experienced less (rather than more) grief than the partner. So, regardless of “objective” differences in levels of grief, over time, it seems that people who perceive their partner as grieving more intensely experience declines in their relationship satisfaction. 3
Figure 2 places the above examples within the DPM-R framework and gives further, testable illustrations of task-specific research across the individual and family levels. These are hypothetical but draw on related empirical evidence and theoretical or clinical insights. 4 A more global approach is conceivable, too, whereby such tasks are included in an overarching test of the model (after all, the original models postulate connectedness between the tasks and an overall effect of working through them somewhat sequentially). One might, for example, address the question, to what extent does intervening along the DPM-R task lines in general improve family and individual functioning? One could explore the role of individual- and family-level tasks following an (adapted) intervention protocol such as that developed for family-focused grief therapy (see Kissane & Bloch, 2002). This program was designed to improve poor family functioning and encourage the sharing of grief among family members (it positively affected individual family members’ outcomes) and as such may be appropriate for testing the DPM-R (e.g., using randomized controlled trial methodology; comparing with a different intervention program).
Examples of integration of loss-oriented and restoration-oriented individual- and family-level tasks using the Dual Process Model–Revised, relevant findings, and directions for future research. Many examples come from bereaved parents’ reactions following the loss of a child, since the suggested constructs derive from previous research on this type of bereavement. Reactions in other relationships are also illustrated (e.g., that of a surviving parent and child); these too merit further expansion.
General Remarks
We return now to Gilbert’s (1989) wisdom, with which we began, to underline the importance of an integrative perspective such as that of the DPM-R: In her qualitative study of couples coping with grief, she reported on—in our terms—a POSR-like phenomenon, noting the protective role of husbands in relation to their wives. Importantly, she commented on the potential negative impact on the partner (and marital relationship): If she were to interpret his behavior as uncaring and cold, rather than loving and protective as he had intended, the result would be a conflict between intent on his part and interpretation on hers . . . this was one of the most common forms of mistaken meanings in behavior. (1989, p. 614)
Gilbert’s identification of such “incongruent grieving” and its effects are systematically featured in the DPM-R (e.g., confrontation vs. nonconfrontation alone and among family members).
We have shown that families make a difference, at both individual and family levels, in (mal)adaptation to bereavement, providing some scientific substantiation for what actually seems like a common-sense position—namely, that families not only contribute to but also help provide solutions for difficulties in bereavement. We have illustrated that the integration of the family with the individual level has incremental value. Therefore, we expect inclusion of the family perspective to increase the power to predict outcomes of bereaved family members. It is apparent, though, that strong empirical support thus far is very limited. Testing the DPM-R remains a challenge: Can it actually be shown that working through the individual and family tasks is really the best road to recovery? Research also needs to establish precisely what roles families play in bereavement outcomes. For example, what types of interpersonal interactions between family members affect (both positively or negatively) the intensity of grief and family functioning in the short and long term?
There is also scope for probing the impact of different family constellations on grief and grieving. For the sake of simplicity, we ignored the diversity that the simple word family covers (cf. Boss & Dahl, 2014). Nor have we systematically addressed different family roles and relationships (or those of other sources of support, as in cases of family absence). Also, families themselves have increased in complexity across the decades, with many now incorporating changed constellations (e.g., divorced parents with children, parents with new partners, stepfamilies, gay parents with children). Nowadays, family members are more geographically widespread; ways of communication are far more diverse than in the past. It is not hard to surmise that bereavement may affect such “fragmented” families in different ways. But we also need to establish if such families are worse or better off in bereavement (at individual and family levels). Again, these are issues for future research to address.
Finally, what are the implications of such an integrated model for professional intervention? Although some family therapy programs have become well established in recent decades (e.g., Kissane & Lichtenthal, 2008; Nadeau, 2008; Sandler, Wolchik, Ayers, Tein, & Luecken, 2013), and Jeffreys (2014) has outlined “healing tasks” for families, the prevailing paradigm is still individual care (cf. Kissane & Parnes, 2014); integrated individual- and family-level models to guide intervention have been lacking. Such a model has been outlined above. However, it remains to be seen whether including family-level tasks within individual programs indeed promotes adjustment among bereaved family members.
Footnotes
Declaration of Conflicting Interests
The authors declared that they had no conflicts of interest with respect to their authorship or the publication of this article.