|
Differentiating Symptoms of Complicated Grief and Depression Among Psychiatric Outpatients
Statistically Significant Change in the Dimensions
The repeated-measures ANOVAs indicated statistically significant change for each of the 5 dimensions over the course of short-term group therapy for CG (Table 2). They also revealed a significant time by treatment effect for the grief symptoms dimension (F = 5.48, df 1, 105, P = 0.021). Improvement in grief symptoms was significantly greater in interpretive therapy than in supportive therapy. Medication was not significantly related to change on any of the 5 dimensions.
|
Table 2 Examination of change on each dimension following short-term group
therapy using repeated-measures ANOVA
|
|
Dimension
|
F
|
df
|
P
|
|
Grief symptoms
|
124.18
|
1, 105
|
< 0.001
|
|
Grief experiences and attitudes
|
52.48
|
1, 105
|
< 0.001
|
|
Grief avoidance
|
83.93
|
1, 104
|
< 0.001
|
|
Depression-cognitive
|
22.56
|
1, 104
|
< 0.001
|
|
Depression-somatic
|
32.85
|
1, 104
|
< 0.001
|
|
ANOVA = analysis of variance
|
Magnitude of Change in the Dimensions
Table 3 presents the effect sizes for each of the 5 dimensions. Effect size was defined as the pretherapy mean minus the posttherapy mean divided by the pretherapy standard deviation. For all cases combined, larger effect sizes were for grief symptoms and grief avoidance, and smaller effect sizes were for the depression dimensions. Also apparent is the sizable difference in effect sizes between the 2 forms of therapy for grief symptoms. The effect size for interpretive therapy was twice as large.
|
Table 3 Effect sizes for the grief and depression dimensions following
short-term group therapy
|
|
Dimension
|
All cases
|
Interpretive cases
|
Supportive cases
|
|
Grief symptoms
|
0.99
|
1.42
|
0.71
|
|
Grief experiences and attitudes
|
0.60
|
0.67
|
0.53
|
|
Grief avoidance
|
1.02
|
1.14
|
0.92
|
|
Depression-cognitive
|
0.42
|
0.53
|
0.31
|
|
Depression-somatic
|
0.60
|
0.58
|
0.65
|
Comparisons among the effect sizes for the 5 dimensions for all therapy completer cases required 10 t-tests. To control for type I error, we used a Bonferroni-adjusted criterion value of 0.005 (0.05/10). The grief symptoms and grief avoidance dimensions did not differ from each other. However, as seen in Table 4, each differed significantly from the grief experiences and attitudes dimension. In addition, the grief symptoms and grief avoidance dimensions differed significantly from each of the depression dimensions. This provides further evidence for the distinction between grief and depression symptoms. Conversely, the grief experiences and attitudes dimension did not differ significantly from either of the depression dimensions. In addition, the 2 depression dimensions did not differ significantly.
|
Table 4 Significant differences in effect sizes among the dimensions for
all therapy completers
|
|
Dimension
|
Compared with
|
t
|
df
|
P
|
|
Grief symptoms
|
Grief experiences and attitudes
|
4.54
|
106
|
< 0.001
|
|
|
Depression-cognitive
|
5.54
|
105
|
< 0.001
|
|
|
Depression-somatic
|
3.49
|
105
|
0.001
|
|
Grief avoidance
|
Grief experiences and attitudes
|
3.43
|
105
|
0.001
|
|
|
Depression-cognitive
|
5.15
|
104
|
< 0.001
|
|
|
Depression-somatic
|
3.23
|
104
|
0.002
|
Discussion
This study found that, among a sample of psychiatric out-patients, CG symptoms emerged as a distinct set of dimensions that were relatively independent of depressive symptoms. The results support findings of previous studies that used nonpatient samples (8,12,13). The present study further differentiated different aspects of CG, highlighting its complex multidimensional nature.
The grief-related dimension that accounted for the most variance reflected many of the symptoms that are most often associated with CG. These include intrusive thoughts and feelings about the lost person, yearning and searching for the lost person, and numbness about the death. We termed these grief symptoms to reflect the prominence of this set of symptoms in CG. The second grief-related dimension, grief experiences and attitudes, consisted of a set of symptoms that conveyed a notion of persistent emotional distress related to the death. The items reflected a propensity to ruminate about the lost person and painful feelings associated with the death. In addition, several items reflected negative attitudes, indicating a lack of willingness to accept the death. Adhering to such attitudes would likely perpetuate the emotional turmoil associated with losing a loved one. The third grief dimension, grief avoidance, reflected an active avoidance of thoughts and feelings associated with the lost person. Symptoms reflecting avoidance have received mixed support in the literature as indicators of CG. However, Horowitz and others (19) view intense avoidance as a key symptom of CG and have suggested that it be one of the criteria for diagnosing CG.
Depressive symptoms loaded on 2 dimensions, independently from any of the grief-related symptoms. The item content of these 2 depression dimensions is remarkably similar to the item content of 2 factors found by Steer and others (20) in their work with the BDI-II. One dimension reflected self-blame and dislike and despair. Steer and colleagues termed this dimension depression-cognitive, which we adopted. The second dimension reflected somatic fatigue and lack of energy. Again, to be consistent with Steer and colleagues, we termed this dimension depression-somatic.
For the most part, the content of each of the 5 dimensions was conceptually meaningful. The dimensions were also internally consistent, as indicated by the high alpha coefficients for each. A high correlation was observed between the grief symptoms and grief experiences and attitudes dimensions. This may be attributed to overlap in the content that many items from each factor address. One possibility as to why the items from the TRIG loaded on a separate factor is that the questionnaire format of the TRIG differed from that used for the 2 other grief measures. The TRIG asked respondents to rate how they presently felt about a person’s death on a 5-point scale. The PGI and IES each asked respondents to rate the extent to which the items were true for them during the past week on a 4-point scale. Further research is required to determine whether these 2 factors truly represent different aspects of CG.
The correlations between each of the 3 grief-related dimensions and each of the 2 depression-related dimensions were fairly low, reflecting considerable independence. This supports previous findings of Prigerson and colleagues (13). Such findings do not imply that the grief and depression dimensions cannot cooccur, rather that each dimension represents a distinct form of emotional dysfunction.
To test the validity of the 5 dimensions, we examined whether the grief and depression dimensions were differentially affected by short-term group therapy for CG. We found that, although all dimensions demonstrated significant change, the greatest change occurred for the grief symptoms and grief avoidance dimensions. There was significantly less change on the grief experiences and attitudes dimension relative to the other 2 grief-related dimensions. The 2 depression dimensions also evidenced significantly smaller amounts of change compared with the grief symptoms and grief avoidance dimensions. The depression dimensions likely demonstrated less change because the therapies provided were focused on helping patients resolve issues related to their losses, not issues related to their depressive symptoms.
It is interesting to note the dissimilarity of the effect size for the grief experiences and attitudes dimension, compared with the other 2 grief dimensions. The effect size for this dimension was more similar to that of the depression dimensions. One possible explanation for why this dimension demonstrated less change is that patients’ attitudes toward their death losses may be a stable characteristic and more difficult to change than grief symptoms and avoidance behaviours. Longer treatment may be required to demonstrate greater change on this dimension. In addition, the forms of group therapy that were provided in this study devoted relatively little attention to modifying negative attitudes toward death losses. Modifying negative attitudes is more consistent with cognitive-behavioural therapy.
We also found that patients who participated in interpretive group therapy had significantly greater improvement on the grief symptoms dimension, compared with patients who participated in supportive group therapy. This suggests that interpretive therapy may be more appropriate to address certain aspects of CG. The greater effectiveness of interpretive group therapy for treating grief symptoms may have been due to its primary objective of helping patients understand and resolve unconscious conflicts that underlie grief symptoms. These unconscious conflicts were assumed to serve as impediments to the patients’ mourning process. Conversely, the primary objective of supportive group therapy was to improve the patients’ immediate adaptation to their life situation through teaching problem-solving skills and guidance. The assumption was that a normal mourning process would occur once an adaptive level of functioning was restored. Thus, in supportive therapy, the conflicts underlying and sustaining grief symptoms were not explored.
There are both limitations and strengths associated with the present study. With regard to limitations, the cross-sectional design of the study does not allow us to determine the stability of the factor structure across time. Future studies with large samples and repeated-measures designs are necessary to determine whether our 5-factor structure is stable. Second, because we chose a select set of grief measures for our study, it is possible that some aspects of complicated grief reactions were not assessed. As a result, our factor structure may not reflect the full spectrum of complicated grief reactions. Similarly, because we used only 1 measure of depression (the BDI), we do not know whether grief and depression symptoms would overlap more if a different measure of depression were used. Third, because PCA is an exploratory factor analytic technique, the results of the present study must be considered preliminary and require replication in a different sample using confirmatory factor analysis. Finally, although we found a significant difference in outcome between the 2 forms of group therapy, it was for only 1 of the 5 outcome dimensions.
With regard to strengths, the present study used a large sample. This provided a favorable subject-to-item ratio of 7:1 for our PCA. The sample also included subjects with mixed ages and types of losses. Thus, our study used a more representative sample than several previous studies that used elderly subjects with spousal losses. Further, the sample consisted of psychiatric outpatients, not community volunteers as in many other studies, thus making our findings more applicable to patient populations. Finally, the present study attempted to validate the dimensions that were identified by examining their responsiveness to group therapy for CG.
Conclusion
The present study illustrated the multidimensional nature of CG and defined some of its components. Clinicians should attend to these different areas of distress when assessing patients who have lost an important person through death. In addition, this study demonstrated the independence of CG and depression symptoms. The independence of CG and depression symptoms was further validated by the finding that they responded differently to short-term group therapy for CG. Clinicians should not assume the absence of CG if depressive symptoms are not present.
The findings of this study also have implications for treatment. Given that an accurate understanding of a patient’s difficulties is a precursor to efficacious treatment, the assessment of the different dimensions of CG and depression is important. Treatments can be tailored to address particular areas of disturbance. The present study suggests that interpretive group therapy may be more effective than supportive group therapy for treating the classic grief symptoms associated with CG.
Funding and Support
This research project was supported by Grant MT-13481 from the Medical Research Council of Canada.
References
1. Stroebe M, von Son M, Stroebe W, Kleber R, Schut H, van den Bout J. On the classification and diagnosis of pathological grief. Clin Psychol Rev 2000;20:57–75.
2. Jacobs S, Mazure C, Prigerson H. Diagnostic criteria for traumatic grief. Death Stud 2000;24:185–99.
3. Rando TA. Treatment of complicated mourning. Champaign (IL): Research Press; 1993.
4. Robinson L, Pickett M. Assessment of adult sibling grief: a review of measurement issues. Hospice J 1996;11:1–18.
5. Hansson RO, Carpenter BN, Fairchild SK. Measurement issues in bereavement. In: Stroebe M, Stroebe W, Hansson RO, editors. Handbook of bereavement. Cambridge (UK): Cambridge University Press; 1993. p 62–74.
6. Clayton PJ. Bereavement and depression. J Clin Psychiatry 1990;51:34–40.
7. Robinson PJ, Fleming S. Differentiating grief and depression. Hospice J 1989;5:77–88.
8. Prigerson H, Bierhals A, Kasl S, Reynolds CF, Shear MK, Newsom JT, and others Complicated grief as a disorder distinct from bereavement-related depression and anxiety: a replication study. Am J Psychiatry 1996;153:1484–6.
9. Piper WE, Ogrodniczuk JS, Azim HF, Weideman R. Prevalence of losses and levels of complicated grief in psychiatric outpatient clinics. Psychiatr Serv 2001;52:1069–74.
10. Piper WE, McCallum M, Joyce AS, Rosie JS, Ogrodniczuk JS. Patient personality and time-limited group psychotherapy for complicated grief. Int J Group Psychother 2001;51:525–52.
11. Zisook S, Lyons L. Bereavement and unresolved grief in psychiatric outpatients. Omega 1989–1990;20:307–22.
12. Prigerson H, Shear K, Newsom J, Frank E, Reynolds CF, Maciejewski PK, and others. Anxiety among widowed elders: is it distinct from depression and grief? Anxiety 1996;2:1–12.
13. Prigerson H, Frank E, Kasl S, Reynolds CF, Anderson B, Zubenko GS, and others. Complicated grief and bereavement-related depression as distinct disorders: preliminary empirical validation in elderly bereaved spouses. Am J Psychiatry 1995;152:22–30.
14. Horowitz MJ, Wilner N, Alvarez W. Impact of Event Scale: a measure of subjective stress. Psychosom Med 1979;41:209–18.
15. Weissman MM, Bothwell S. Assessment of social adjustment by patient self-report. Arch Gen Psychiatry 1976;33:1111–5.
16. Faschingbauer TR, Zisook S, DeVaul RA. The Texas Revised Inventory of Grief. In: Zisook S, editor. Biopsychosocial aspects of bereavement. Washington (DC): American Psychiatric Press; 1987. p 111–24.
17. Beck AT, Steer RA, Brown GK. Manual for Beck Depression Inventory-II. San Antonio (TX): Psychological Corporation; 1996.
18. Cattell RB. The scree test for the number of factors. Multivariate Behav Res 1966;1:245–76.
19. Horowitz MJ, Siegal B, Holen A, Bonanno GA, Milbrath C, Stinson CH. Diagnostic criteria for complicated grief disorder. Am J Psychiatry 1997;154:904–10.
20. Steer RA, Ball R, Ranieri WF. Dimensions of the Beck Depression Inventory-II in clinically depressed outpatients. J Clin Psychol 1999;55:117–28.
Authors
1. Assistant Professor, Department of Psychiatry, University of British Columbia, Vancouver, British Columbia.
2. Professor, Department of Psychiatry, University of British Columbia, Vancouver, British Columbia.
3. Associate Professor, Department of Psychiatry, University of Alberta, Edmonton, Alberta
4. Clinical Assistant Professor, Department of Psychiatry, University of British Columbia, Vancouver, British Columbia.
5. Clinical Professor, Department of Psychiatry, University of Alberta, Edmonton, Alberta.
6. Clinical Professor, Department of Psychiatry, University of British Columbia, Vancouver, British Columbia.
7. Clinical Professor, Department of Psychiatry, University of Alberta, Edmonton, Alberta.
Address for correspondence: Dr J Ogrodniczuk, 2250 Wesbrook Mall, University of British Columbia, Vancouver, BC V6T 1W6
e-mail: ogrodnic@interchange.ubc.ca
1 | 2 | 3
|
