| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
| ||||||||||||||||||||||||
RESEARCH ARTICLE |
1 Lighthouse International, New York.
2 Department of Psychology, State University of New York at Stony Brook.
3 Department of Counseling and Clinical Psychology, Teachers College, Columbia University, New York.
Address correspondence to Kathrin Boerner, Lighthouse International, 111 East 59th Street, New York, NY 10022. E-mail: kboerner{at}lighthouse.org
 |
Abstract |
|---|
 TOP Abstract MethodsResultsDiscussionReferences |
|---|
IN THE United States alone, more than 900,000 people are widowed each year, and nearly three fourths of these are over the age of 65 years (Moss, Moss, & Hansson, 2001
; U.S. Department of Human Services, 1996). Because the social networks of elderly people are often narrowed by deteriorating health and death, spousal relationships often assume increasing importance. Over the course of a marriage, many couples form a highly interdependent relationship that is based on interlocking roles, commitments, and traditions (Moss et al., 2001). Elderly couples may have such closely interwoven lives that the death of one partner may cut across the very meaning of the other's existence (Raphael, 1983). Moreover, elderly people are far more likely than their younger counterparts to remain in a widowed state (Wilson & Clarke, 1992).
Coming to terms with personal loss is an important feature of successful aging (Baltes & Skrotzki, 1995). Despite the prevalence of conjugal bereavement among the elderly population and the centrality of this loss, there has been a paucity of research examining reactions to widowhood among this group (Lund, 1989), and most of the available research has focused on the first 2 years of bereavement. It is generally believed that, following the loss of a spouse, elders experience considerable distress, which will decrease over time as they come to terms with the loss (Malkinson, 2001). However, there is growing recognition that there are marked individual differences in how elders respond to conjugal loss (Lund, 1989; Bonanno et al., 2002). Some elders show relatively little distress during the first several months following their spouse's death, whereas others show considerable distress.
Our major purpose in this study was to examine what happens to these groups over time. Are those who show relatively little distress in the initial months after the loss best understood as resilient, or are such individuals simply denying grief symptomatology that will emerge at a later date? Conversely, will those who show consistently high distress following the loss continue to manifest such distress indefinitely, or will their emotional pain subside over time? These questions have critically important implications for subsequent theory and intervention work with the bereaved. For example, if it were common for those who manifest little distress after the loss to show a "delayed" grief reaction, then such findings would underscore the importance of preventive interventions that might foster "working through" the loss in a more timely manner.
It is commonly assumed by both lay persons and bereavement theorists that people who respond to the loss of a spouse with relatively little distress are not coping with the loss appropriately. For example, Bowlby (1980) considered the "prolonged absence of conscious grieving" (p. 138) to be a form of disordered mourning. Osterweis, Solomon, and Green (1984) noted that clinicians commonly assume that "the absence of grieving phenomena following bereavement represents some form of personality pathology" (p. 18). Many investigators assume that "absent grief" is indicative of denial and that bereaved individuals who fail to show overt grief will experience a delayed grief response (e.g., Middleton, Moylan, Raphael, Burnett, & Martinek, 1993; also see Jacobs, 1993). Rando (1993) posited that the absence of grief is one of the strongest indicators that complicated mourning will ensue, and that treatment may be necessary. Specifically, she suggested that therapists should challenge a bereaved person's denial, explore why the mourner feels unable to accept the death, and address his or her fear that mourning will be overwhelming. Another interpretation in such a case is that the person was never that attached to the deceased (e.g., Fraley & Shaver, 1999), perhaps because of a general tendency to remain emotionally distant (Rando, 1993) or because the relationship was "purely narcissistic with little recognition of the real person who was lost." (Raphael, 1983, pp. 205–206).
In contrast to these assumptions, empirical evidence has increasingly demonstrated that a substantial number of bereaved individuals do not appear to experience intense distress beyond the initial weeks after a loss (Bonanno, Wortman, & Nesse, 2004). These studies also indicate that delayed grief is exceedingly rare (see Bonanno & Field, 2001 for a review). However, as we already noted, most of this research has focused on the first 2 years following the loss. Rando (1993) pointed out that mourning "may be delayed for many years," and that "delayed mourning is indistinguishable from absent or inhibited mourning until the delay is over" (p. 159; see also Worden, 2002). Lund (1989) noted that this is particularly the case among elderly individuals, who experience the death of their spouse against the backdrop of other age-related changes such as chronic illness, disability, and diminished physical and cognitive capacities. Interestingly, research has found that, over the first 2 years, the emotional distress associated with grief declines more slowly among older persons than younger persons (Sanders, 1981). Hence, it is important to follow the bereavement process beyond the common 2-year time frame.
Another variant from the norm that has been identified as problematic is chronic grief. In this type of situation, grieving does not seem to progress normally; instead, mourners become "stuck" in their grief and continue to show high distress. If this is the case, it may be anticipated that chronic grief will continue indefinitely (Rando, 1993). However, it is unclear how long such reactions last, particularly among elderly respondents, and whether they eventually resolve the loss on their own. As Jacobs (1993) emphasized, understanding how bereaved people heal, and how the intensity of pain subsides over time, is one of the most important and least well-understood research questions with regard to grief.
Thus, it is important to determine how bereaved persons who show consistently low distress for the first several months after the loss, as well as those who go through a period of intense distress, chronic grief, or depression, ultimately adjust to this loss. Moreover, judging the degree of health or pathology in response to loss typically involves assumptions about the way the loss has been processed. Therefore, a long-term assessment should also include an examination of how the bereaved are processing the loss. For example, the relative absence of grief symptoms can be taken to more clearly reflect positive adaptation if also accompanied by reports of a low need to avoid thoughts about the deceased. Our purpose in the present study is to examine extended patterns of outcome among individuals who had initially exhibited either stable low distress or chronic grief during the initial months of bereavement. In an earlier study, researchers defined grief trajectories by using data from 3 years preloss through 18 months postloss based on a sample of older adults (Bonanno et al., 2002). Here we discuss an examination of these patterns over a 4-year period in terms of both bereaved elders' adaptation and processing of the loss.
Delineating Patterns of Bereavement
In addition to the two distinct reactions to loss described herein, that is, chronic grief and the relative absence of grief, or resilience, Bonanno and colleagues (2002) delineated three other core outcome trajectories. They assigned one such trajectory, referred to as the common grief pattern, to participants with low levels of depression prior to the loss who had a grief reaction at 6 months, but whose depression scores at 18 months postloss were not different from their preloss level of depression. They assigned another core trajectory, the improvement in depression following conjugal loss, to participants with elevated depression prior to the loss that improved markedly after the loss and remained at low levels of depression through 18 months. The improved pattern has received relatively little research attention. However, when the marital relationship is characterized by a high degree of conflict, or by a stressful caregiving experience, the death of a loved one may result in improved well-being for the bereaved. Under these conditions, the spouse's death may provide relief and hence represent the end of a chronic stressor rather than a stressor per se (Wheaton, 1990). Evidence in support of this notion comes from a recent study on loss from Alzheimer's disease (Schulz et al., 2003). Finally, another trajectory that has also received little attention is chronic depression (i.e., elevated depression levels both prior to and after the loss for an extend period of time), which Bonanno and colleagues assigned to participants with elevated depression both prior to and after their spouses' death. More detailed operational definitions of these five trajectories are provided in the Methods section.
To further characterize the nature of these patterns, Bonanno and associates (2002) identified their preloss predictors. Resilience was associated with acceptance of death and belief in a just world. By contrast, the depressed-improved individuals had the most negative and ambivalent preloss evaluations of their marriage, and they were likely to have had an ill spouse. In addition, they were likely to be emotionally unstable and to believe that the world was generally unjust to them. Chronic grievers were likely to have had healthy spouses, to rate their marriage very positively, and to show high levels of preloss dependency (e.g., agreeing that no one could take the spouse's place). The chronically depressed group was less positive about their marriage than chronic grievers, but just as dependent on their spouses.
Further analyses examined the context and processing of the loss at 6 and 18 months postloss (Bonnano et al., 2004). Results suggest that chronic grief stems from an enduring struggle with cognitive and emotional distress related to the loss, whereas chronic depression results more from enduring emotional difficulties that are exacerbated by the loss. Both the resilient and depressed-improved groups showed healthy profiles and relatively little evidence of either struggling with or avoiding the loss during bereavement. Finally, the resilient group reported deriving the most comfort from positive memories, whereas the chronically depressed reported deriving the least. This is important, because it indicates that resilient people are indeed positively attached to their deceased spouses.
The Current Study
In the current study we examined these same trajectories over a longer time period: from 18 to 48 months postloss. Addressing the question of long-term consequences may hold critical implications for intervention by helping to identify the processes most likely to be associated with long-term difficulties. It is also important to identify those patterns of grief that are not likely to portend mental health problems, to minimize the possibility that those who adjust well to their loss will not be unnecessarily pathologized (Bonanno & Kaltman, 2001).
One major research question of interest was whether or not the resilient stay resilient and the depressed-improved remain improved in the long term, 4 years following the loss of their spouse. The assumption that the absence of acute distress is based on avoidance and denial (e.g., Rando, 1993) suggests that respondents in these groups should show efforts to avoid the loss (e.g., seeking distraction, trying not to think about the deceased or the loss), and that grief or depression would emerge over time (i.e., delayed grief). In contrast, drawing on the prior research just summarized, we predicted that participants in the resilient group would continue to do well at 48 months postloss, meaning that they would neither show delayed grief nor evidence a lack of cognitive and emotional involvement with the deceased.
We also suspected that the depressed-improved group might do less well than the resilient group at Month 48 compared with Month 18. This suspicion is based on the research just reported, which demonstrated that, before the loss, this group showed features (e.g., emotional instability) with a potential to reemerge and cause difficulties in the long run. For the common grievers, we expected depression to remain low. Considering that this group reflects a pattern of going from low to high distress, which then abated by the time of the 18-month follow-up, we found it likely that respondents would remain at this level.
We further investigated whether the chronic grievers and the chronically depressed would remain distressed up to 48 months postloss. As already noted, most classic grief theorists (e.g., Jacobs, 1993) discuss the notion of chronic grief but fail to indicate how long it typically takes and how often it abates. We assumed that the distress of the chronic grief group may persist longer than in the other trajectory groups, because respondents experienced a more devastating loss (i.e., losing a beloved, healthy spouse, on whom they depended). We expected that, although these preloss factors may result in more persistent grief, this group would show at least some improvement from Month 18 to Month 48. In contrast, because the chronically depressed group showed a relatively stable pattern of depression that was present even before the loss, it seemed likely to us that something besides the loss was contributing to their depression. Therefore, we expected their depression to remain high longer than that in the other groups.
|
METHODS |
|---|
|
TOPAbstractMethodsResultsDiscussionReferences |
|---|
Participants from the CLOC study who subsequently lost a spouse were identified using daily obituaries in three Detroit-area newspapers and monthly death record tapes provided by the State of Michigan. The National Death Index was used to confirm deaths and obtain information about cause of death. Widowed respondents were asked to participate in follow-up interviews at 6, 18, and 48 months after their spouses' deaths. Of the 319 respondents who participated in the baseline interview and lost a spouse, 86% (n = 276) participated in at least one and 64% (n = 205) participated in at least two follow-up interviews. The primary reason for nonresponse was ill health or death at follow-up (42%) and refusal to participate (38%).
The subsample (n = 92) we used in the present analysis included those respondents on whom we had data for all three follow-up time points and who had been assigned to the grief trajectory groups identified in prior research on the CLOC data (Bonanno et al., 2002). Respondents were 90% women, and the average age was70 years (SD = 6.2). This subsample did not differ from those without data for the last follow-up time point with regard to age, gender, income, self-rated overall health, history of clinical depression, and depressive symptoms at baseline. They did, however, differ in terms of educational level (t = 2.8, p <.01) and depressive symptoms at 6 as well as at 18 months postloss (t = –2.5, p <.05 and t = –2.9, p <.01, respectively). Respondents in the present subsample were likely to have a higher educational level and to report fewer depressive symptoms at the 6- and 18-month follow-up. In particular, the latter difference is of concern with respect to the interpretation of findings for those bereaved participants who showed elevated levels of depression initially and at the short-term follow-up (i.e., the chronic grief and chronic depression group; see the subsequent group definitions). We address possible implications for the study findings in the Discussion section.
Defining Grief Trajectory Groups
The initial grief trajectories were identical to those reported by Bonanno and colleagues (2002). They defined these trajectories by using longitudinal data from the Center for Epidemiologic Studies Depression (CES-D) scale (Radloff, 1977) obtained an average of 3 years prior to the spouses' deaths (M = 36.7 months, SD = 16.6 months) and at approximately 6 and 18 months after the spouses' deaths. They followed a three-step procedure. First, they categorized participants as having either high or low preloss depression by using the 80th percentile as a cutoff for high depression. Second, they calculated two change scores for each participant by comparing CES-D scores at preloss with the 6-month follow-up, and with the 18-month follow-up. They categorized each change score as either a grief reaction, if depression increased relative to preloss by 1 SD or greater; improved functioning, if depression decreased by greater than 1 SD; or no change, if depression scores remained constant or increased or decreased by less than 1 SD.
To accommodate the possibility that depression scores might have decreased or increased in part because of regression to the mean, they defined change separately for the high and low preloss depression groups by using the standard deviation of each group. In addition, because preloss depression scores tended to cluster around the sample mean, they assigned a grief reaction only when depression scores during bereavement increased to greater than the 50th percentile for the larger sample. In a third step, they combined the two change scores to create eight possible bereavement outcome patterns. They used only the patterns that were exhibited by at least 5% of the sample. These patterns encompassed 185 participants, or 90.2% of the sample.
Of the present study sample (N = 92), including those participants who had data on all three follow-up time points and had been assigned to the grief trajectory groups as defined herein, group sample sizes were as follows: common grief, n = 10 (10.9%); depressed–improved, n = 17 (18.5%); resilient, n = 46 (50%); chronic grief, n = 10 (10.9%); and chronic depression, n = 9 (9.8%).
Adjustment
We measured depressive symptoms with the 11-item CES-D scale (Iowa form; Kohut, Berkman, Evans, & Cornoni-Huntley, 1993), a reduced version of the original 20-item CES-D (Radloff, 1977). Respondents indicated on a 3-point Likert-type scale that ranged from (0) hardly ever to (2) most of the time how often they experienced the symptoms in the past week (higher score = higher depressive symptomatology; 
=.85).
We measured grief by using 16 items derived from the Bereavement Index (Jacobs, Kasl, & Ostfeld, 1986), the Present Feelings About Loss scale (Singh & Raphael, 1981) and the Texas Revised Inventory of Grief (Zisook, DeVaul, & Click, 1982). We derived the total grief score from the average of six subscales (anger, shock, yearning, intrusive thoughts, anxiety, and despair), each scored on a 4-point scale (1 = never; 2 = rarely; 3 = sometimes; 4 = often) pertaining to the past month ( =.88).
Processing the Loss Variables
We measured the extent of thinking about the deceased by using a one-item indicator: During the past month, how often have you had thoughts or memories of your husband or wife?
We measured seeking distraction by averaging four items: During the past month, have you tried to keep busy so that you would be less likely to dwell on your husband or wife or his or her death? I try not to think about what happened. To cope with these feelings, how much have you gotten out of the house, for example, gone somewhere by talking a walk or a drive? How much have you kept busy or tried to get involved in some activity? ( =.60).
We measured "searching for meaning" and "asking why me" by a single question: During the past month, have you ever found yourself searching to make sense of or find some meaning in your husband or wife's death? During the past 12 months, did you ever ask yourself, "Why me?" or "Why my husband or wife?" Would you say no, never; yes, but rarely; yes, sometimes; or yes, often?
|
RESULTS |
|---|
|
TOPAbstractMethodsResultsDiscussionReferences |
|---|
|
As we predicted, the resilient group generally showed the most positive picture in terms of adjustment compared with the other groups (see Figure 1 for depression over time; similar pattern for grief not shown). As we expected, the common grief displayed a fairly stable pattern over time with regard to both outcomes. To test the hypothesis that the depressed-improved group would show poorer adjustment than the resilient group at 48 months postloss, we conducted planned contrasts comparing the two groups on both outcomes at this final time point. The depressed-improved group had significantly higher scores for depressive symptoms, contrast t(17.62) = –1.87, p <.05, and grief, contrast t(87) = –1.77, p <.05.
|
Findings also supported the prediction that both the resilient and the common grief groups would continue to do well up to 48 months postloss. Thus, there seemed to be no indication of delayed grief for the resilient group. Although we had considered that the depressed-improved group might develop adjustment problems over time, there was only a trend in this direction with regard to depression scores; t(17) = 1.82, p <.10. However, group comparisons reflect average differences and do not reveal information about individual cases. Therefore, the following set of analyses examined individual patterns of grief and depression in the depressed-improved and resilient groups.
Elevated and delayed grief
The most widely used cutoff point for clinically relevant levels of depression has been set at the 80th percentile (e.g., Beekman et al., 2002; Comstock & Helsing, 1976). We used the 80th percentile based on the baseline CLOC sample of N = 1,532 to extract a cutoff point that would reflect depression levels in the general aging population. With a scale range of 0–22, the so-determined cutoff score was 7 (3–4 out of 11 symptoms in the past week). For grief, we defined elevated scores on the basis of the median of grief scores at 6 months postloss (Mdn = 2; scale range 1–4), with scores above the median indicating elevated grief. This cutoff score also reflected the 80th percentile of grief levels at 48 months postloss. As in Bonnano and Field (2001), we operationalized delayed grief as not having elevated scores at 6 months but instead at the long-term follow-up (Month 48).
Using this operationalization, we found that delayed grief did not emerge for respondents in either of the two groups. Similarly, we could not find delayed depressive symptoms for participants in the resilient group. Although two participants from the depressed-improved group had depression scores above 7 at 48 months postloss, these scores reflected only a slight elevation (8 and 9) relative to the cutoff point (7). Thus, there was no systematic evidence for delayed depression or grief in either of the two groups.
Processing the loss
There were only two significant group effects for the processing the loss variables (searching for meaning, and asking why me; see Table 1). Significant time effects emerged for seeking distraction, searching for meaning, and extent of thinking about the spouse. The latter variable also showed a significant Group x Time effect, reflecting that the chronically depressed group showed the largest decrease over time in extent of thinking about the deceased between 18 and 48 months postloss, compared with the other groups, who showed smaller decreases (resilient, chronic grief,) or fairly stable scores (common, depressed-improved). However, at 48 months postloss, the chronic depression group's mean score (M = 4.1) for this variable was comparable with the total group mean score. Across all five groups, by this time point, thinking about the deceased (M = 4.4, SD = 1.3) appeared to be generally more prevalent than the effort of trying not to think about him or her (seeking distraction: M = 1.4, SD =.77). Most importantly, the continuously positive adjustment on both outcome measures for the resilient and one outcome for the depressed-improved group did not seem to be based on a processing approach that was characterized by blocking out thoughts about the deceased.
We conducted planned contrasts weighted for the prediction that the chronically depressed group would show more difficulties in processing the loss at the 48-month follow-up compared with chronic grievers. As we expected, these contrast analyses showed that, at 48 months postloss, the chronic depression group had a significantly higher score on searching for meaning, contrast t(87) = –2.86, p <.01, and on asking why me, contrast t(87) = –1.73, p <.05, than the chronic grief group. In general, the chronically depressed group reported the highest levels of asking why me and of searching for meaning at both Month 18 and 48. This suggests that participants in this group were most likely to be searching for the meaning of the loss and thinking about why this had happened to them at 48 months following the loss.
|
DISCUSSION |
|---|
|
TOPAbstractMethodsResultsDiscussionReferences |
|---|
These findings add to the growing body of evidence challenging the notion of delayed grief as a likely consequence of the failure to become intensely distressed following the loss of a loved one. However, because the evidence for a continuously positive adjustment was not as consistent for the depressed-improved group as it was for the resilient group, future research should further explore the possibility of delayed adaptational problems among bereaved elders who show improvement following their spouse's death. Earlier, we noted that very little research has been devoted to the depressed-improved pattern, largely because most bereavement studies did not have sufficient data to identify such a group. As a result of societal trends, it is becoming increasingly common for elders to lose a spouse following a chronic and debilitating illness, and to be called on to perform caregiving activities while they are frail and in poor health themselves (Carr, 2003). Hence, it will behoove researchers in future studies to utilize prospective designs that might adequately capture this pattern within the context of caregiving (see, e.g., Schulz et al., 2003). Utilizing such a design, researchers may find it possible to identify the conditions under which subsequent depression or existential questions about meaning are particularly likely to emerge. One hypothesis suggested by our findings is that subsequent depression may emerge after a period of well-being in those cases in which people are required to perform stressful caregiving duties for a spouse for whom they have negative or ambivalent feelings.
The differential findings regarding the chronic grief and chronically depressed group also underscored the need to further refine the criteria that are used to identify those who may be at risk for long-term problems. The chronic grief group experienced a more intense and prolonged period of distress compared with, for example, the common grief group, which does not seem surprising if preloss characteristics (e.g., losing a beloved spouse who was healthy) are considered. Improvements on outcome and processing the loss measures, however, indicated a turn toward better adjustment by the 48-month time point, which suggests that this group does not remain chronically distressed as a result of the loss.
In contrast to the chronic grief group, the chronically depressed group clearly demonstrated long-term problems, with little indication of improvement between Month 18 and 48. This group not only showed the poorest adjustment 4 years after the loss but also struggled the most with questions about meaning. One possible explanation for the latter finding is that those who are chronically depressed are prone to ruminate about their situation (Nolen-Hoeksema, Parker, & Larson, 1994), and that searching for meaning of the loss and asking why me is part of this more general rumination tendency. Alternately, such a tendency may lie at the core of the depression and explain its chronicity. Future research that examines the interrelationships between specific responses such as searching for meaning and asking why me, more general rumination tendencies, and depression among bereaved elders may help clarify this issue.
Two limitations that should be acknowledged are the issue of attrition and small sample size, common problems in the investigation of long-term adaptation. First, the relatively small size of some subgroups caused by attrition reduced statistical power and may have limited our ability to detect possible effects. Second, although a significant portion of the sample had incomplete longitudinal data because the study ended, some of the attrition was due to respondents' dropping out over the course of the study. Considering this attrition pattern, we find it important to acknowledge that participants who had data through 48 months were less depressed than participants compared with those who did not. This means that the percentage of respondents with poor long-term adaptation may have been higher if these respondents had been retained in the study. Thus, our findings should be considered with caution because some of the respondents in the chronic grief group who did not have data for the last time point may have continued to show a chronic pattern. Moreover, although our conclusions regarding the long-term problems and possible treatment needs of the chronically depressed group might even be more strongly founded as a result of this attrition bias, both the findings for the chronic grief and chronic depression group have to be replicated. However, the resilient and depressed-improved groups were unlikely to be affected by this attrition pattern.
In sum, the findings presented in this article provide further support for the major conclusion that emerged from the prior findings (Bonanno et al., 2002, 2004); "doing well" after a loss is not necessarily a cause for concern but rather a normal response for many older adults. The differential findings for the chronic grief and chronic depression group underscore the need to further refine the criteria that are used to identify those who are at risk for long-term problems. The present findings tentatively suggest that bereaved elders who show a trajectory of chronic depression might benefit from a different intervention focus than those with a chronic grief pattern. Individuals with chronic depression may be helped most by interventions that focus on their enduring emotional problems, and that assist them in dealing with the day-to-day strains associated with widowhood. If this approach is not successful, pharmacological interventions may also be beneficial for bereaved persons with chronic depression (Zisook & Shuchter, 2001). Chronic grievers, in contrast, may benefit from an intervention that acknowledges the centrality of their loss and helps them in processing the loss, in developing a new identity, and in restructuring their life so that it gradually becomes more meaningful and rewarding (Neimeyer, 2000).
|
Acknowledgments |
|---|
|
Footnotes |
|---|
Received for publication April 1, 2004. Accepted for publication September 16, 2004.
|
References |
|---|
|
TOPAbstractMethodsResultsDiscussionReferences |
|---|
| ||||||||||||||||||||||||
| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
