Clinical Psychology and Psychotherapy Clin. Psychol. Psychother. 23, 118–124 (2016) Published online in 22 December 2014 Wiley Online Library (wileyonlinelibrary.com). DOI: 10.1002/cpp.1940
Exploring Outcomes Related to Anxiety and Depression in Completers of a Randomized Controlled Trial of Complicated Grief Treatment Kim Glickman,1* M. Katherine Shear2 and Melanie Wall3 1
Social Work, York College/CUNY, Jamaica, NY, USA School of Social Work, Columbia University, New York, NY, USA 3 Psychiatry, Columbia University, New York, NY, USA 2
The present study examines a more fine-grained analysis of anxiety-related and depression-related outcomes amongst a sample of treatment completers who were assigned to complicated grief treatment (CGT) (n = 35) or interpersonal psychotherapy (IPT) (n = 34) in a previously reported randomized controlled trial. We examined effects of antidepressant use and measures of anxiety and depression, focusing especially on guilt related to the death or deceased and grief-related avoidance in order to further understand the differential effectiveness of CGT and IPT amongst participants who received the full course of treatment. Analyses showed that CGT produced greater reductions in anxiety and depressive symptoms including negative thoughts about the future and grief-related avoidance. CGT’s advantage over IPT in lowering depression was most pronounced amongst those not taking antidepressants. Our results further elucidate CGT effects and support the idea that CG and major depressive disorder are distinct conditions. Copyright © 2014 John Wiley & Sons, Ltd. Key Practitioner Message: • Targeted treatment for complicated grief (CG) produces benefits in associated mood and anxiety symptoms and CG symptoms. • Amongst patients with CG, interpersonal psychotherapy seems relatively ineffective in ameliorating depressive symptoms. • Grief-related depressive symptoms may not respond to standard treatments unless CG symptoms are also addressed. • Reducing grief-related symptoms, such as anxieties about the future, guilt related to the death or deceased and avoidance of reminders of the loss may be important aspects in reducing CG. Keywords: Complicated Grief, Complicated Grief Treatment, Interpersonal Psychotherapy, Grief-related Depressive Symptoms, Avoidance Behaviour, Grief-related Guilt
The loss of a loved one is one of the most difficult experiences a person can endure. Whilst most move successfully through a normal grief process, roughly 7% of bereaved people develop a debilitating condition variously known as complicated grief (CG) (Kersting, 2011), prolonged grief disorder (Prigerson et al., 2009) or persistent complex bereavement disorder (Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition). The syndrome is characterized by persistent yearning and longing, intense sorrow and emotional pain and pre-occupation with thoughts of the loved one (Maercker et al., 2013). We published the first randomized controlled treatment study targeting symptoms of CG (Shear et al., 2005) reporting *Correspondence to: Kim Glickman, College/CUNY, Jamaica, NY, USA. E-mail: [email protected]
Social
Copyright © 2014 John Wiley & Sons, Ltd.
Work,
York
significantly greater improvement with a targeted CG treatment (CGT) compared with interpersonal psychotherapy (IPT). Depression and impairment were also significantly more improved with CGT amongst treatment completers (Shear et al., 2005). Simon et al (2008) found that antidepressant medication use was significantly associated with treatment completion rates (91% versus 58%) for CGT but not IPT (70% versus 77%) and marginally better outcomes for the sample as a whole. Two other CGtargeted treatment studies have shown improvement in general psychopathology (Boelen et al., 2007; Wagner et al., 2006) and grief-related negative cognitions and avoidance behaviour (Boelen et al., 2011). The purpose of the present study is to explore differential treatment effects on anxiety-related and depressionrelated outcomes amongst participants in our previously reported study who received a full course of treatment.
Exploring Outcomes Related to Anxiety and Depression in CGT We were interested in a more fine-grained analysis of these symptoms and especially in examining those symptoms that comprise grief complications that are seen as interfering with the natural grief process and specifically targeted in the CGT model. These include negative thoughts about the future, guilt related to the death or deceased and griefrelated avoidance. CGT specifically targets counterfactual thinking around troubling aspects of the death (e.g., selfblame for how the person died), avoidance behaviour and problems with emotion regulation. We hypothesized that these symptoms would respond better to CGT than to IPT. Given previous findings of differential effects of antidepressants in the two treatment groups, we further explored possible moderator effects of antidepressant use.
METHOD Details of the methodology of the parent study are available in a previously published paper (Shear et al., 2005). Briefly, participants were recruited through professional referral, self-referral and media announcements between 2001 and 2004 in Pittsburgh, PA at a university-based psychiatric clinic and a satellite clinic in a low-income African-American community. Participants were eligible if they had lost a loved one at least 6 months prior and met criteria for CG, defined as a score of 30 or higher on the Inventory of CG (ICG) (Prigerson et al., 1995). Participants were randomly assigned to receive IPT or CGT and were offered 16 weekly therapy sessions over a 16 to 20 week period. Medication for longer than 3 months at a stable dose for >6 weeks was permitted. Data were collected by self-report and independent evaluators blinded to treatment assignment at baseline, during treatment and post-treatment follow-up. The current study includes treatment completers (n = 69, 35 in CGT and 34 in IPT), a subset of study participants (n = 95, CGT = 49; IPT = 46). We focus on completers in order to examine these outcomes amongst those who actually received each treatment. Completers did not differ from dropouts on any of the baseline measures. The control condition, IPT, is a manualized short-term treatment that focuses on a selected problem area of grief, interpersonal disputes, role transition or interpersonal as deficits. Most of the IPT treatments in this study focused on grief alone or grief and one of the other problem areas. The experimental condition, CGT, was also delivered according to a treatment manual and had yielded positive results in a prior pilot study (Shear et al., 2001). CGT is based on an attachment theory model of grief. It integrates IPT with cognitive–behavioural techniques used to address trauma-like symptoms of CG and motivational interviewing techniques and incorporates the dual process model of coping (Stroebe & Schut, 1999). Unlike IPT, CGT includes grief monitoring; psychoeducation about CG; Copyright © 2014 John Wiley & Sons, Ltd.
119 loss-focused procedures using exposure-based techniques such as imaginal revisiting (repeated retelling of the story of the death); and situational revisiting (graded exposure to avoided situations), structured memories forms and an imaginal conversation with the deceased. Each CGT session also includes restoration-focused techniques such as work with personal aspirations and goals, building support and planning rewarding activities.
Measures
Anxiety Symptoms Anxiety was assessed with the Beck Anxiety Inventory (BAI), a 21-item, well-validated instrument used to measure clinical anxiety with high internal consistency (α = 0.92) (Beck et al., 1988). We examined cognitive and somatic subscales of the BAI (Hewitt & Norton, 1993) to examine different types of anxiety symptoms separately.
Depressive Symptoms Depressive symptoms were measured using the Beck Depression Inventory (BDI) (Beck et al., 1979) and the Hamilton Rating Scale for Depression (HRSD) (Hamilton, 1960). The BDI is a widely used, well-validated 21-item self-report instrument used to assess the severity of depression. In the current analysis, we examine cognitive and somatic–affective subscales (Steer et al., 1999) and changes in item Q2 related to negative thoughts about the future. This item is targeted in CGT by the aspirational goals component of CGT. Results for the total scale were previously reported (Shear et al 2005). Depression was also measured by the HRSD, a widely used, wellvalidated 25-item interview-rated instrument used to assess the severity of depression (Hamilton, 1960).
Changes in Complicating Symptoms of Avoidance and Dysfunctional Thinking The CGT of model (Shear, 2012) conceptualizes CG as comprised of prolonged acute grief symptoms and complicating cognitive, behavioural and emotional symptoms. Avoidance behaviour and second-guessing cognitions, such as self-blaming, are two important grief complications. These constructs were measured by instruments developed for our studies. The Grief-related Avoidance Questionnaire (GRAQ) (Shear et al., 2007) is a 15-item self-report questionnaire that rates avoidance of common grief-related situations and activities. The scale has good psychometric properties (α = 0.87), and CG patients endorse a range of scores on the scale. Clin. Psychol. Psychother. 23, 118–124 (2016)
K. Glickman et al.
120 Self-blame related to the deceased was assessed using three items from the Structured Clinical Interview of CG (SCI-CG). The SCI-CG is a 30-item semi-structured clinical instrument used to assess symptoms of CG. The intraclass correlation coefficient for test–retest reliability (n = 218) was 0.68, 95% CI [0.60, 0.75] (Bui et al., 2014). The three items used to assess self-blame related to the deceased included the following: have you had guilty or self-blaming thoughts or beliefs about the death? Do you blame yourself for doing or not doing something either when _____ was alive, or at the time he or she died, that you think might have helped? Do you have the idea that you could have prevented this death, even though you know it is not very rational? (Cronbach’s α = 0.86).
dividing the mean change (baseline to post-treatment) by the pooled baseline standard deviation to facilitate comparison across different outcomes (referred to in Table 3 as ‘standardized change’). The Wilks–Shapiro test was used to check change variables for large departures from normality. Where found, sensitivity analysis was carried out to assess the robustness of the p-values by additionally performing the Wilcoxon signed-rank test. To test for moderation of effects by antidepressant use, multiple regression analysis was used to examine the interaction between antidepressant use and treatment group as it affects change in associated outcomes. Twosided tests were used for assessing the interaction (as we did not have strong prior directionality for our hypothesis) with a p-value of 0.10 for statistical significance because of the difficulty of detecting interaction effects.
Statistical Analysis Differences in baseline characteristics between treatment groups were tested using two-sample t-tests. We hypothesized that improvements in all outcomes would be greater in the CGT group compared with the IPT group; hence, one-tailed t-tests were used to test for improvements from baseline to week 16. We are aware that this method can inflate the type-1 error rate; however, in all previous analyses from this study, CGT was more effective than IPT (Shear et al., 2005). Statistical significance was set at p < =0.05. Effect size (ES) for the treatment differences between groups was calculated by dividing the mean difference in outcome (baseline to post-treatment) by the pooled baseline standard deviation. We also examined change within each treatment group using paired t-tests. Mean within-treatment change was standardized by Table 1.
RESULTS Baseline Characteristics Table 1 includes demographic characteristics of the completer sample by group. Overall, the sample was 84% women, 73% Caucasian, 22% African-American and 3% Asian, with a mean age of 50 years. Mean ICG score was 44.8. Comorbid concurrent mood or anxiety disorders were common. Forty-six percent had major depressive disorder (MDD), 52% post-traumatic stress disorder, 17% generalized anxiety disorder, 10% panic disorder, 3% social phobia disorder and 6% obsessive–compulsive disorder. Thirty-three out of 69 samples (48%) were taking antidepressant medication. There were no significant
Baseline comparison of treatment groups (n = 69)
Variables Sample characteristics Men White Age, mean (SD) Antidepressant use (at least one) Major depressive disorder Post-traumatic stress disorder Generalized anxiety disorder Panic disorder Social phobia Obsessive–compulsive disorder Inventory of Complicated Grief, mean (SD) Work and Social Adjustment Scale, mean (SD) Beck Depression Inventory, mean (SD) Beck Anxiety Inventory, mean (SD) Hamilton Rating Scale for Depression, mean (SD)
CGT (n = 35)a
IPT (n = 34)a
p-value
5 (14) 24 (69) 52 (13.4) 19 (54) 18 (53)b 19 (56)b 6 (18) 5 (14) 1 (3)b 4 (11)b 46.1 (8.5) 21.9 (10.8) 25.6 (10.9) 17.7 (12.6) 19.1 (7.0)
6 (18) 26 (77) 48 (11.4) 14 (41) 13 (38) 16 (47) 6 (18) 2 (6) 1 (3) 0 (0) 43.4 (9.8) 20.5 (9.6) 21.8 (9.9) 14.5 (9.8) 16.8 (6.8)
0.75 0.59 0.20 0.15 0.33 0.47 1.0 0.43 1.0 0.11 0.21 0.56 0.14 0.25 0.18
CGT = complicated grief treatment; IPT = interpersonal psychotherapy; SD = standard deviation. a Data are expressed as no. (%) unless otherwise noted. b n = 34 due to missing data
Copyright © 2014 John Wiley & Sons, Ltd.
Clin. Psychol. Psychother. 23, 118–124 (2016)
Exploring Outcomes Related to Anxiety and Depression in CGT
121
Symptom Change by Treatment Group
differences between the two randomized groups at baseline on any measure.
Changes in Anxiety Symptoms Cognitive symptoms assessed on the BAI decreased significantly in both treatment groups and greater in CGT than IPT (1.01 standardized change in CGT versus 0.56 standardized change in IPT, with a medium ES difference 0.45, p = 0.02). Somatic symptoms of anxiety (BAI) decreased modestly, but results were similar for both treatment groups (0.42 in CGT versus 0.43 in IPT) (Figure 1 and Table 3).
Antidepressant Use as a Moderator Table 2 shows the resulting summary of pre-treatment and post-treatment scores on the HRSD by treatment group and antidepressant status. Amongst eight outcomes we examined, one showed an interaction between antidepressant use and treatment group on HRSD (b = 6.88, t(65) = 1.9, p = 0.06) with greater difference between CG and IPT for individuals who were not on antidepressants. Amongst participants not on antidepressants, CGT reduced depression but IPT did not (1.6 standardized change in CGT versus 0.03 in IPT, p = 0.0001). For those taking medication, both groups had a reduction in depression scores with a marginally greater change in CGT (0.93 in CGT versus 0.50 in IPT, p = 0.13).
Changes in Depressive Symptoms Depressive symptoms on the HRSD decreased significantly with CGT but not IPT (1.2 standardized change in CGT versus 0.21 standardized change in IPT), leading to a large ES difference (ES = 0.99, p = 0.0002). Cognitive symptoms of depression on the BDI decreased signifi-
Table 2. Pre-treatment and post-treatment scores on Hamilton Rating Scale for Depression by treatment group and antidepressant status No antidepressants CGT (n = 14) M (SD) Pre-treatment Post-treatment Difference
18.6 (6.1) 8.7 (5.0) 9.9 (6.7)
Antidepressants
IPT (n = 20)
p-value
M (SD)
0.0000
16.3 (6.2) 16.5 (9.8) 0.2 (6.6)
p-value
0.55
CGT (n = 19) M (SD) 19.8 (7.1) 12.9 (9.8) 6.9 (8.2)
IPT (n = 14)
p-value
M (SD)
0.0009
17.6 (7.8) 13.9 (6.3) 3.7 (7.6)
p-value
0.046
CGT = complicated grief treatment; IPT = interpersonal psychotherapy; SD = standard deviation. Note: n = 34 in the no-antidepressant group and 33 in the antidepressant group rather than 35 and 34, respectively, due to missing data.
Figure 1. Change in symptoms by treatment group
Copyright © 2014 John Wiley & Sons, Ltd.
Clin. Psychol. Psychother. 23, 118–124 (2016)
K. Glickman et al.
122 Table 3.
Pre-treatment and post-treatment scores by group and effect sizes for treatment differences
Variables Anxiety symptoms Cognitive (BAI) Pre-treatment Post-treatment Change Standardized changec p-value of standardized change Somatic (BAI) Pre-treatment Post-treatment Difference Standardized change p-value of standardized change Depressive symptoms HRSD Pre-treatment Post-treatment Change Standardized change p-value of standardized change Cognitive (BDI) Pre-treatment Post-treatment Change Standardized change p-value of standardized change Somatic–affective (BDI) Pre-treatment Post-treatment Change Standardized change p-value of standardized change Negative thoughts Pre-treatment Post-treatment Change Standardized change p-value of standardized change Grief-related symptoms Avoidance – GRAQ Pre-treatment Post-treatment Difference Standardized change p-value of standardized change Guilt/self-blame – SCI-CG Pre-treatment Post-treatment Change Standardized change p-value of standardized change
CGT Mean (SD)
IPT Mean (SD)
n = 35 7.8 (4.4) 3.5 (4.0) 4.3 (3.3) 1.01 0.00 n = 35 6.2 (5.8) 4.0 (6.7) 2.2 (3.8) 0.42 0.00
n = 34 6.4 (4.1) 4 (4.1) 2.4 (4.5) 0.56 0.00 n = 34 6.0 (5.1) 3.7 (4.4) 2.3 (3.9) 0.43 0.00
n = 35 19.3 (6.6) 11.2 (8.3) 8.1 (7.6) 1.2 0.00 n = 35 7.9 (5.5) 4.4 (4.8) 3.5 (3.7) 0.66 0.00 n = 35 9.4 (4.7) 5.6 (4.9) 3.8 (4.3) 0.82 0.00 n = 35 1.4 (0.91) 0.51 (0.66) 0.89 (0.96) 1.0 0.00
n = 34 16.8 (6.8) 15.4 (8.5) 1.4 (7.2) 0.21 0.13 n = 34 7.4 (5.1) 4.8 (5.4) 2.6 (3.4) 0.49 0.00 n = 34 8.1 (4.6) 6.0 (4.5) 2.1 (3.5) 0.46 0.00 n = 34 1.1 (0.83) 0.68 (0.81) 0.41 (0.61) 0.47 0.00
n = 29 25.5 (11.1) 16.9 (12.6) 8.6 (14.9) 0.72 0.00 n = 25 4.5 (2.2) 1.9 (1.9) 2.6 (2.1) 1.17 0.00
n = 25 21.4 (12.8) 18.5 (15.6) 3.0 (9) 0.25 0.06 n = 31 4.4 (2.2) 2.7 (2.5) 1.7 (2.6) 0.77 0.00
Effect sizea
p-valueb
0.45
0.02
0.01
0.53
0.99
0.00
0.17
0.15
0.36
0.04
0.54
0.01
0.47
0.05
0.40
0.09
CGT = complicated grief treatment; IPT = interpersonal psychotherapy; SD = standard deviation; BAI = Beck Anxiety Inventory; HRSD = Hamilton Rating Scale for Depression; BDI = Beck Depression Inventory; SCI-CG = Structured Clinical Interview of Complicated Grief. a Effect size of treatment is calculated following Cohen’s D, i.e., (CGT mean Diff IPT mean Diff)/pooled baseline standard deviation of measure, e.g., 2 2 for guilt/self-blame (SCI-CG), we have (2.6 1.7)/sqrt((2.2) + 30 × (2.2) )/(24 + 30)) = 0.40. b p-value for one-sided test of treatment effect, specifically the test for a difference in differences, i.e., Ho: change CGT = change IPT versus Ha: change CGT >change IPT. c Standardized change values for outcomes are plotted in Figure 1.
Copyright © 2014 John Wiley & Sons, Ltd.
Clin. Psychol. Psychother. 23, 118–124 (2016)
Exploring Outcomes Related to Anxiety and Depression in CGT cantly in both groups (0.66 in CGT versus 0.49 in IPT), and the differential effect was not significant (ES = 0.17, p = 0.15). Somatic–affective symptoms on the BDI also decreased in both groups (0.82 in CGT versus 0.46 in IPT) with an ES difference (ES = 0.36, p = 0.04) (Figure 1 and Table 3). Negative thoughts about the future (Q2) also decreased significantly in both treatment groups (1.0 standardized change in CGT versus 0.47 in IPT), and the medium ES difference was statistically significant (ES = 0.54, p = 0.01).
Changes in Complicating Symptoms of Avoidance and Dysfunctional Thinking Avoidance as measured by the GRAQ total score decreased significantly in CGT (0.72) but not IPT (0.25) with a medium ES difference (ES = 0.47, p = 0.05) (Figure 1 and Table 3). Guilt/self-blame related to the death measured by the SCI-CG decreased significantly in both groups with a medium ES difference that was not statistically significant (1.17 in CGT versus 0.77 in IPT, ES = 0.40, p = 0.09).
DISCUSSION This study extends research by Shear et al. (2005) by providing information for treatment completers on differential effects of CGT and IPT for anxiety and depression. We found greater changes with CGT than IPT in anxiety, depression, grief-related avoidance and negative thoughts about the future. These results confirm relatively low responsiveness of individuals with CG to IPT even amongst those who received a full course of treatment. Our findings indicate that a CG-targeted treatment produces unique benefits for people with CG in mood and anxiety as well as grief symptoms. We found statistically and clinically significant reduction in cognitive symptoms of anxiety and depression, including negative thoughts about the future, and in avoidance of reminders of the loss. These results are consistent with the CGT model, which posits that complicating thoughts, emotions and behaviours interfere with the normal grieving process and form a part of the syndrome of CG. According to this model, negative cognitions activated in CG distract the person from focusing on the finality and consequences of the loss (Shear, 2012). Negative thoughts about the future can keep a person with CG stuck in a constant state of worry about bad things that might happen and a conviction that life can never be good again. Avoidance behaviour also interferes with information processing necessary for successful adaptation to an important loss. Copyright © 2014 John Wiley & Sons, Ltd.
123 Our results further show strikingly greater effects of CGT than IPT on depressive symptom severity as measured by the Hamilton Depression ratings. Reduction in HRSD scores was almost 10 times greater for CGT than IPT. Notably, this difference was more pronounced amongst those not taking medication. Amongst participants who received CGT, reduction is HRSD occurred whether or not they were already taking antidepressant medication with the reduction slightly greater for those not on antidepressants. By contrast, participants treated with IPT showed no improvement in HRSD score unless they were also taking antidepressants. Even still, reduction in HRSD was nearly twice as great for CGT compared with IPT. These results strongly support the idea that CG and MDD are distinct conditions. We also examined depression results as measured by the BDI. We previously reported that amongst completers CGT showed a significantly greater reduction than IPT in BDI scores (Shear et al 2005). In the current study, we found this difference was accounted for by a reduction in somatic/affective symptoms. Cognitive and somatic anxiety symptoms as measured by the BAI showed improvements in both CGT and IPT. We previously reported marginally better outcome for CGT versus IPT completers on the BAI total score. In the present study, cognitive symptoms of anxiety showed statistically significantly better results for CGT, whereas the somatic subscale of the BAI showed no difference between groups. This study has several limitations. By design, we were interested in focusing on change in symptoms amongst those who received the assigned treatment, so our sample included only completers. Although an analysis of baseline variables showed no difference between completers and dropouts on any measure, we did not have information on change in associated symptoms for those who discontinued treatment, so we do not know if results would be the same for an intent-to-treat sample. Additionally, our sample size of 69 is relatively small, and power to detect differences between the treatment groups is limited. Moreover, two of the instruments we used in our analyses (SCI-CG and GRAQ) were developed after we began the study, and the sample size for the variables created from these scales was somewhat lower (n = 54). This study of anxiety and depression outcomes in treatment completers shows that CGT was more effective than IPT in targeting these associated symptoms, possibly because CGT specifically targets dysfunctional thoughts and avoidance behaviours as manifestations of grief complications. Additionally, IPT has been repeatedly shown to be an efficacious treatment for depression and our results confirm the importance of distinguishing CG from MDD. Clin. Psychol. Psychother. 23, 118–124 (2016)
124
REFERENCES Beck, A., Epstein, N., Brown, G., & Steer, R. (1988). An inventory for measuring clinical anxiety: Psychometric properties. Journal of Clinical Psychiatry, 56(6), 893–897. Beck, A., Rush, A., Shawb,F., & Emery,G. (1979). Cognitive therapy of depression. New York: Guilford. Boelen, P., de Keijser, J., van den Hout, M., & van den Bout, J. (2007). Treatment of complicated grief: A comparison between cognitive–behavioral therapy and supportive counseling. Journal of Consulting and Clinical Psychology, 75(2), 277–284. Boelen, P., de Keijser, J., van den Hout, M., & van den Bout, J. (2011). Factors associated with outcome of cognitive– behavioral therapy for complicated grief: A preliminary study. Clinical Psychology and Psychotherapy, 18, 284–291. Bui, E., Robinaugh, D., Mauro, C., Wang, Y., Zeng, D., Duan, N., Reynolds, C., Zisook, S., Lebowitz, B., Simon, N. & Shear, M.K. (2014). Development and preliminary psychometric properties for a Structured Clinical Interview for Complicated Grief. Paper presented at the 30th Annual Meeting of the International Society for Traumatic Stress Studies (poster presentation), Miami, FL. Hamilton, M. (1960). A rating scale for depression. Journal of Neurology, Neurosurgery and Psychiatry, 25, 56–67. Hewitt, P., & Norton, G. (1993). The Beck Anxiety Inventory: A psychometric analysis. Psychological Assessment, 5(4), 408–412. Kersting, A. (2011). Prevalence of complicated grief in a representative population-based sample. Journal of Affective Disorders, 131, 339–343. Maercker, A., Brewin, C. R., Bryant, R. A., Cloitre, M., Reed, G. M., van Ommeren, M., et al. (2013). Proposals for mental disorders specifically associated with stress in the International
Copyright © 2014 John Wiley & Sons, Ltd.
K. Glickman et al. Classification of Diseases-11. Lancet, 381(9878), 1683–1685. doi: 10.1016/S0140-6736(12)62191-6 Prigerson, H., Bierhals, A., Stanislav, K., Reynolds III, C., Shear, K., Day, N., et al. (1995). Inventory of Complicated Grief: A scale to measure maladaptive symptoms of loss. Psychiatry Research, 59, 65–79. Prigerson, H., Horowitz, M., Jacobs, S., Parks, C., Asian, M., Goodkin, K., et al. (2009). Prolonged grief disorder: Psychometric validation of criteria proposed for DSM-V and ICD-11. PLoS Medicine, 6, 1–12. Shear, K. (2012). Grief and mourning gone awry: Pathway and course of complicated grief. Dialogues in Clinical Neuroscience, 14(2), 119–128. Shear, K., Frank, E., Foa, E., Cherry, E., Reynolds III, C., Vander Bilt, J., et al. (2001). Traumatic grief treatment: A pilot study. American Journal of Psychiatry, 158(9), 1506–1508. Shear, K., Frank, E., Houck, P., & Reynolds III, C. (2005). Treatment of complicated grief: A randomized controlled trial. JAMA, 293(21), 2601–2608. Shear, K., Monk, T., Houck, P., Melhem, N., Frank, E., Reynolds, C., et al. (2007). An attachment-based model of complicated grief including the role of avoidance. European Archives of Psychiatry and Clinical Neuroscience, 257(8), 453–461. Simon, N. M., Shear, K., Fagiolini, A., Frank, E., Zalta, A., Thompson, E., Reynolds III, C., Silowash, R. . (2008). Impact of concurrent naturalistic pharmacotherapy on psychotherapy of complicated grief. Psychiatry Research, 159(1-2), 31. Steer, R. A., Ball, R., & Ranieri, W. (1999). Dimensions of the Beck Depression Inventory-II in clinically depressed outpatients. Journal of Clinical Psychology, 55(1), 117–128. Wagner, B., Knaevelsrud, C., & Maercker, A. (2006). Internetbased cognitive-behavioral therapy for complicated grief: A randomized controlled trial. Death Studies, 30, 429–453.
Clin. Psychol. Psychother. 23, 118–124 (2016)
Exploring Outcomes Related to Anxiety and Depression in Completers of a Randomized Controlled Trial of Complicated Grief Treatment Kim Glickman,1* M. Katherine Shear2 and Melanie Wall3 1
Social Work, York College/CUNY, Jamaica, NY, USA School of Social Work, Columbia University, New York, NY, USA 3 Psychiatry, Columbia University, New York, NY, USA 2
The present study examines a more fine-grained analysis of anxiety-related and depression-related outcomes amongst a sample of treatment completers who were assigned to complicated grief treatment (CGT) (n = 35) or interpersonal psychotherapy (IPT) (n = 34) in a previously reported randomized controlled trial. We examined effects of antidepressant use and measures of anxiety and depression, focusing especially on guilt related to the death or deceased and grief-related avoidance in order to further understand the differential effectiveness of CGT and IPT amongst participants who received the full course of treatment. Analyses showed that CGT produced greater reductions in anxiety and depressive symptoms including negative thoughts about the future and grief-related avoidance. CGT’s advantage over IPT in lowering depression was most pronounced amongst those not taking antidepressants. Our results further elucidate CGT effects and support the idea that CG and major depressive disorder are distinct conditions. Copyright © 2014 John Wiley & Sons, Ltd. Key Practitioner Message: • Targeted treatment for complicated grief (CG) produces benefits in associated mood and anxiety symptoms and CG symptoms. • Amongst patients with CG, interpersonal psychotherapy seems relatively ineffective in ameliorating depressive symptoms. • Grief-related depressive symptoms may not respond to standard treatments unless CG symptoms are also addressed. • Reducing grief-related symptoms, such as anxieties about the future, guilt related to the death or deceased and avoidance of reminders of the loss may be important aspects in reducing CG. Keywords: Complicated Grief, Complicated Grief Treatment, Interpersonal Psychotherapy, Grief-related Depressive Symptoms, Avoidance Behaviour, Grief-related Guilt
The loss of a loved one is one of the most difficult experiences a person can endure. Whilst most move successfully through a normal grief process, roughly 7% of bereaved people develop a debilitating condition variously known as complicated grief (CG) (Kersting, 2011), prolonged grief disorder (Prigerson et al., 2009) or persistent complex bereavement disorder (Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition). The syndrome is characterized by persistent yearning and longing, intense sorrow and emotional pain and pre-occupation with thoughts of the loved one (Maercker et al., 2013). We published the first randomized controlled treatment study targeting symptoms of CG (Shear et al., 2005) reporting *Correspondence to: Kim Glickman, College/CUNY, Jamaica, NY, USA. E-mail: [email protected]
Social
Copyright © 2014 John Wiley & Sons, Ltd.
Work,
York
significantly greater improvement with a targeted CG treatment (CGT) compared with interpersonal psychotherapy (IPT). Depression and impairment were also significantly more improved with CGT amongst treatment completers (Shear et al., 2005). Simon et al (2008) found that antidepressant medication use was significantly associated with treatment completion rates (91% versus 58%) for CGT but not IPT (70% versus 77%) and marginally better outcomes for the sample as a whole. Two other CGtargeted treatment studies have shown improvement in general psychopathology (Boelen et al., 2007; Wagner et al., 2006) and grief-related negative cognitions and avoidance behaviour (Boelen et al., 2011). The purpose of the present study is to explore differential treatment effects on anxiety-related and depressionrelated outcomes amongst participants in our previously reported study who received a full course of treatment.
Exploring Outcomes Related to Anxiety and Depression in CGT We were interested in a more fine-grained analysis of these symptoms and especially in examining those symptoms that comprise grief complications that are seen as interfering with the natural grief process and specifically targeted in the CGT model. These include negative thoughts about the future, guilt related to the death or deceased and griefrelated avoidance. CGT specifically targets counterfactual thinking around troubling aspects of the death (e.g., selfblame for how the person died), avoidance behaviour and problems with emotion regulation. We hypothesized that these symptoms would respond better to CGT than to IPT. Given previous findings of differential effects of antidepressants in the two treatment groups, we further explored possible moderator effects of antidepressant use.
METHOD Details of the methodology of the parent study are available in a previously published paper (Shear et al., 2005). Briefly, participants were recruited through professional referral, self-referral and media announcements between 2001 and 2004 in Pittsburgh, PA at a university-based psychiatric clinic and a satellite clinic in a low-income African-American community. Participants were eligible if they had lost a loved one at least 6 months prior and met criteria for CG, defined as a score of 30 or higher on the Inventory of CG (ICG) (Prigerson et al., 1995). Participants were randomly assigned to receive IPT or CGT and were offered 16 weekly therapy sessions over a 16 to 20 week period. Medication for longer than 3 months at a stable dose for >6 weeks was permitted. Data were collected by self-report and independent evaluators blinded to treatment assignment at baseline, during treatment and post-treatment follow-up. The current study includes treatment completers (n = 69, 35 in CGT and 34 in IPT), a subset of study participants (n = 95, CGT = 49; IPT = 46). We focus on completers in order to examine these outcomes amongst those who actually received each treatment. Completers did not differ from dropouts on any of the baseline measures. The control condition, IPT, is a manualized short-term treatment that focuses on a selected problem area of grief, interpersonal disputes, role transition or interpersonal as deficits. Most of the IPT treatments in this study focused on grief alone or grief and one of the other problem areas. The experimental condition, CGT, was also delivered according to a treatment manual and had yielded positive results in a prior pilot study (Shear et al., 2001). CGT is based on an attachment theory model of grief. It integrates IPT with cognitive–behavioural techniques used to address trauma-like symptoms of CG and motivational interviewing techniques and incorporates the dual process model of coping (Stroebe & Schut, 1999). Unlike IPT, CGT includes grief monitoring; psychoeducation about CG; Copyright © 2014 John Wiley & Sons, Ltd.
119 loss-focused procedures using exposure-based techniques such as imaginal revisiting (repeated retelling of the story of the death); and situational revisiting (graded exposure to avoided situations), structured memories forms and an imaginal conversation with the deceased. Each CGT session also includes restoration-focused techniques such as work with personal aspirations and goals, building support and planning rewarding activities.
Measures
Anxiety Symptoms Anxiety was assessed with the Beck Anxiety Inventory (BAI), a 21-item, well-validated instrument used to measure clinical anxiety with high internal consistency (α = 0.92) (Beck et al., 1988). We examined cognitive and somatic subscales of the BAI (Hewitt & Norton, 1993) to examine different types of anxiety symptoms separately.
Depressive Symptoms Depressive symptoms were measured using the Beck Depression Inventory (BDI) (Beck et al., 1979) and the Hamilton Rating Scale for Depression (HRSD) (Hamilton, 1960). The BDI is a widely used, well-validated 21-item self-report instrument used to assess the severity of depression. In the current analysis, we examine cognitive and somatic–affective subscales (Steer et al., 1999) and changes in item Q2 related to negative thoughts about the future. This item is targeted in CGT by the aspirational goals component of CGT. Results for the total scale were previously reported (Shear et al 2005). Depression was also measured by the HRSD, a widely used, wellvalidated 25-item interview-rated instrument used to assess the severity of depression (Hamilton, 1960).
Changes in Complicating Symptoms of Avoidance and Dysfunctional Thinking The CGT of model (Shear, 2012) conceptualizes CG as comprised of prolonged acute grief symptoms and complicating cognitive, behavioural and emotional symptoms. Avoidance behaviour and second-guessing cognitions, such as self-blaming, are two important grief complications. These constructs were measured by instruments developed for our studies. The Grief-related Avoidance Questionnaire (GRAQ) (Shear et al., 2007) is a 15-item self-report questionnaire that rates avoidance of common grief-related situations and activities. The scale has good psychometric properties (α = 0.87), and CG patients endorse a range of scores on the scale. Clin. Psychol. Psychother. 23, 118–124 (2016)
K. Glickman et al.
120 Self-blame related to the deceased was assessed using three items from the Structured Clinical Interview of CG (SCI-CG). The SCI-CG is a 30-item semi-structured clinical instrument used to assess symptoms of CG. The intraclass correlation coefficient for test–retest reliability (n = 218) was 0.68, 95% CI [0.60, 0.75] (Bui et al., 2014). The three items used to assess self-blame related to the deceased included the following: have you had guilty or self-blaming thoughts or beliefs about the death? Do you blame yourself for doing or not doing something either when _____ was alive, or at the time he or she died, that you think might have helped? Do you have the idea that you could have prevented this death, even though you know it is not very rational? (Cronbach’s α = 0.86).
dividing the mean change (baseline to post-treatment) by the pooled baseline standard deviation to facilitate comparison across different outcomes (referred to in Table 3 as ‘standardized change’). The Wilks–Shapiro test was used to check change variables for large departures from normality. Where found, sensitivity analysis was carried out to assess the robustness of the p-values by additionally performing the Wilcoxon signed-rank test. To test for moderation of effects by antidepressant use, multiple regression analysis was used to examine the interaction between antidepressant use and treatment group as it affects change in associated outcomes. Twosided tests were used for assessing the interaction (as we did not have strong prior directionality for our hypothesis) with a p-value of 0.10 for statistical significance because of the difficulty of detecting interaction effects.
Statistical Analysis Differences in baseline characteristics between treatment groups were tested using two-sample t-tests. We hypothesized that improvements in all outcomes would be greater in the CGT group compared with the IPT group; hence, one-tailed t-tests were used to test for improvements from baseline to week 16. We are aware that this method can inflate the type-1 error rate; however, in all previous analyses from this study, CGT was more effective than IPT (Shear et al., 2005). Statistical significance was set at p < =0.05. Effect size (ES) for the treatment differences between groups was calculated by dividing the mean difference in outcome (baseline to post-treatment) by the pooled baseline standard deviation. We also examined change within each treatment group using paired t-tests. Mean within-treatment change was standardized by Table 1.
RESULTS Baseline Characteristics Table 1 includes demographic characteristics of the completer sample by group. Overall, the sample was 84% women, 73% Caucasian, 22% African-American and 3% Asian, with a mean age of 50 years. Mean ICG score was 44.8. Comorbid concurrent mood or anxiety disorders were common. Forty-six percent had major depressive disorder (MDD), 52% post-traumatic stress disorder, 17% generalized anxiety disorder, 10% panic disorder, 3% social phobia disorder and 6% obsessive–compulsive disorder. Thirty-three out of 69 samples (48%) were taking antidepressant medication. There were no significant
Baseline comparison of treatment groups (n = 69)
Variables Sample characteristics Men White Age, mean (SD) Antidepressant use (at least one) Major depressive disorder Post-traumatic stress disorder Generalized anxiety disorder Panic disorder Social phobia Obsessive–compulsive disorder Inventory of Complicated Grief, mean (SD) Work and Social Adjustment Scale, mean (SD) Beck Depression Inventory, mean (SD) Beck Anxiety Inventory, mean (SD) Hamilton Rating Scale for Depression, mean (SD)
CGT (n = 35)a
IPT (n = 34)a
p-value
5 (14) 24 (69) 52 (13.4) 19 (54) 18 (53)b 19 (56)b 6 (18) 5 (14) 1 (3)b 4 (11)b 46.1 (8.5) 21.9 (10.8) 25.6 (10.9) 17.7 (12.6) 19.1 (7.0)
6 (18) 26 (77) 48 (11.4) 14 (41) 13 (38) 16 (47) 6 (18) 2 (6) 1 (3) 0 (0) 43.4 (9.8) 20.5 (9.6) 21.8 (9.9) 14.5 (9.8) 16.8 (6.8)
0.75 0.59 0.20 0.15 0.33 0.47 1.0 0.43 1.0 0.11 0.21 0.56 0.14 0.25 0.18
CGT = complicated grief treatment; IPT = interpersonal psychotherapy; SD = standard deviation. a Data are expressed as no. (%) unless otherwise noted. b n = 34 due to missing data
Copyright © 2014 John Wiley & Sons, Ltd.
Clin. Psychol. Psychother. 23, 118–124 (2016)
Exploring Outcomes Related to Anxiety and Depression in CGT
121
Symptom Change by Treatment Group
differences between the two randomized groups at baseline on any measure.
Changes in Anxiety Symptoms Cognitive symptoms assessed on the BAI decreased significantly in both treatment groups and greater in CGT than IPT (1.01 standardized change in CGT versus 0.56 standardized change in IPT, with a medium ES difference 0.45, p = 0.02). Somatic symptoms of anxiety (BAI) decreased modestly, but results were similar for both treatment groups (0.42 in CGT versus 0.43 in IPT) (Figure 1 and Table 3).
Antidepressant Use as a Moderator Table 2 shows the resulting summary of pre-treatment and post-treatment scores on the HRSD by treatment group and antidepressant status. Amongst eight outcomes we examined, one showed an interaction between antidepressant use and treatment group on HRSD (b = 6.88, t(65) = 1.9, p = 0.06) with greater difference between CG and IPT for individuals who were not on antidepressants. Amongst participants not on antidepressants, CGT reduced depression but IPT did not (1.6 standardized change in CGT versus 0.03 in IPT, p = 0.0001). For those taking medication, both groups had a reduction in depression scores with a marginally greater change in CGT (0.93 in CGT versus 0.50 in IPT, p = 0.13).
Changes in Depressive Symptoms Depressive symptoms on the HRSD decreased significantly with CGT but not IPT (1.2 standardized change in CGT versus 0.21 standardized change in IPT), leading to a large ES difference (ES = 0.99, p = 0.0002). Cognitive symptoms of depression on the BDI decreased signifi-
Table 2. Pre-treatment and post-treatment scores on Hamilton Rating Scale for Depression by treatment group and antidepressant status No antidepressants CGT (n = 14) M (SD) Pre-treatment Post-treatment Difference
18.6 (6.1) 8.7 (5.0) 9.9 (6.7)
Antidepressants
IPT (n = 20)
p-value
M (SD)
0.0000
16.3 (6.2) 16.5 (9.8) 0.2 (6.6)
p-value
0.55
CGT (n = 19) M (SD) 19.8 (7.1) 12.9 (9.8) 6.9 (8.2)
IPT (n = 14)
p-value
M (SD)
0.0009
17.6 (7.8) 13.9 (6.3) 3.7 (7.6)
p-value
0.046
CGT = complicated grief treatment; IPT = interpersonal psychotherapy; SD = standard deviation. Note: n = 34 in the no-antidepressant group and 33 in the antidepressant group rather than 35 and 34, respectively, due to missing data.
Figure 1. Change in symptoms by treatment group
Copyright © 2014 John Wiley & Sons, Ltd.
Clin. Psychol. Psychother. 23, 118–124 (2016)
K. Glickman et al.
122 Table 3.
Pre-treatment and post-treatment scores by group and effect sizes for treatment differences
Variables Anxiety symptoms Cognitive (BAI) Pre-treatment Post-treatment Change Standardized changec p-value of standardized change Somatic (BAI) Pre-treatment Post-treatment Difference Standardized change p-value of standardized change Depressive symptoms HRSD Pre-treatment Post-treatment Change Standardized change p-value of standardized change Cognitive (BDI) Pre-treatment Post-treatment Change Standardized change p-value of standardized change Somatic–affective (BDI) Pre-treatment Post-treatment Change Standardized change p-value of standardized change Negative thoughts Pre-treatment Post-treatment Change Standardized change p-value of standardized change Grief-related symptoms Avoidance – GRAQ Pre-treatment Post-treatment Difference Standardized change p-value of standardized change Guilt/self-blame – SCI-CG Pre-treatment Post-treatment Change Standardized change p-value of standardized change
CGT Mean (SD)
IPT Mean (SD)
n = 35 7.8 (4.4) 3.5 (4.0) 4.3 (3.3) 1.01 0.00 n = 35 6.2 (5.8) 4.0 (6.7) 2.2 (3.8) 0.42 0.00
n = 34 6.4 (4.1) 4 (4.1) 2.4 (4.5) 0.56 0.00 n = 34 6.0 (5.1) 3.7 (4.4) 2.3 (3.9) 0.43 0.00
n = 35 19.3 (6.6) 11.2 (8.3) 8.1 (7.6) 1.2 0.00 n = 35 7.9 (5.5) 4.4 (4.8) 3.5 (3.7) 0.66 0.00 n = 35 9.4 (4.7) 5.6 (4.9) 3.8 (4.3) 0.82 0.00 n = 35 1.4 (0.91) 0.51 (0.66) 0.89 (0.96) 1.0 0.00
n = 34 16.8 (6.8) 15.4 (8.5) 1.4 (7.2) 0.21 0.13 n = 34 7.4 (5.1) 4.8 (5.4) 2.6 (3.4) 0.49 0.00 n = 34 8.1 (4.6) 6.0 (4.5) 2.1 (3.5) 0.46 0.00 n = 34 1.1 (0.83) 0.68 (0.81) 0.41 (0.61) 0.47 0.00
n = 29 25.5 (11.1) 16.9 (12.6) 8.6 (14.9) 0.72 0.00 n = 25 4.5 (2.2) 1.9 (1.9) 2.6 (2.1) 1.17 0.00
n = 25 21.4 (12.8) 18.5 (15.6) 3.0 (9) 0.25 0.06 n = 31 4.4 (2.2) 2.7 (2.5) 1.7 (2.6) 0.77 0.00
Effect sizea
p-valueb
0.45
0.02
0.01
0.53
0.99
0.00
0.17
0.15
0.36
0.04
0.54
0.01
0.47
0.05
0.40
0.09
CGT = complicated grief treatment; IPT = interpersonal psychotherapy; SD = standard deviation; BAI = Beck Anxiety Inventory; HRSD = Hamilton Rating Scale for Depression; BDI = Beck Depression Inventory; SCI-CG = Structured Clinical Interview of Complicated Grief. a Effect size of treatment is calculated following Cohen’s D, i.e., (CGT mean Diff IPT mean Diff)/pooled baseline standard deviation of measure, e.g., 2 2 for guilt/self-blame (SCI-CG), we have (2.6 1.7)/sqrt((2.2) + 30 × (2.2) )/(24 + 30)) = 0.40. b p-value for one-sided test of treatment effect, specifically the test for a difference in differences, i.e., Ho: change CGT = change IPT versus Ha: change CGT >change IPT. c Standardized change values for outcomes are plotted in Figure 1.
Copyright © 2014 John Wiley & Sons, Ltd.
Clin. Psychol. Psychother. 23, 118–124 (2016)
Exploring Outcomes Related to Anxiety and Depression in CGT cantly in both groups (0.66 in CGT versus 0.49 in IPT), and the differential effect was not significant (ES = 0.17, p = 0.15). Somatic–affective symptoms on the BDI also decreased in both groups (0.82 in CGT versus 0.46 in IPT) with an ES difference (ES = 0.36, p = 0.04) (Figure 1 and Table 3). Negative thoughts about the future (Q2) also decreased significantly in both treatment groups (1.0 standardized change in CGT versus 0.47 in IPT), and the medium ES difference was statistically significant (ES = 0.54, p = 0.01).
Changes in Complicating Symptoms of Avoidance and Dysfunctional Thinking Avoidance as measured by the GRAQ total score decreased significantly in CGT (0.72) but not IPT (0.25) with a medium ES difference (ES = 0.47, p = 0.05) (Figure 1 and Table 3). Guilt/self-blame related to the death measured by the SCI-CG decreased significantly in both groups with a medium ES difference that was not statistically significant (1.17 in CGT versus 0.77 in IPT, ES = 0.40, p = 0.09).
DISCUSSION This study extends research by Shear et al. (2005) by providing information for treatment completers on differential effects of CGT and IPT for anxiety and depression. We found greater changes with CGT than IPT in anxiety, depression, grief-related avoidance and negative thoughts about the future. These results confirm relatively low responsiveness of individuals with CG to IPT even amongst those who received a full course of treatment. Our findings indicate that a CG-targeted treatment produces unique benefits for people with CG in mood and anxiety as well as grief symptoms. We found statistically and clinically significant reduction in cognitive symptoms of anxiety and depression, including negative thoughts about the future, and in avoidance of reminders of the loss. These results are consistent with the CGT model, which posits that complicating thoughts, emotions and behaviours interfere with the normal grieving process and form a part of the syndrome of CG. According to this model, negative cognitions activated in CG distract the person from focusing on the finality and consequences of the loss (Shear, 2012). Negative thoughts about the future can keep a person with CG stuck in a constant state of worry about bad things that might happen and a conviction that life can never be good again. Avoidance behaviour also interferes with information processing necessary for successful adaptation to an important loss. Copyright © 2014 John Wiley & Sons, Ltd.
123 Our results further show strikingly greater effects of CGT than IPT on depressive symptom severity as measured by the Hamilton Depression ratings. Reduction in HRSD scores was almost 10 times greater for CGT than IPT. Notably, this difference was more pronounced amongst those not taking medication. Amongst participants who received CGT, reduction is HRSD occurred whether or not they were already taking antidepressant medication with the reduction slightly greater for those not on antidepressants. By contrast, participants treated with IPT showed no improvement in HRSD score unless they were also taking antidepressants. Even still, reduction in HRSD was nearly twice as great for CGT compared with IPT. These results strongly support the idea that CG and MDD are distinct conditions. We also examined depression results as measured by the BDI. We previously reported that amongst completers CGT showed a significantly greater reduction than IPT in BDI scores (Shear et al 2005). In the current study, we found this difference was accounted for by a reduction in somatic/affective symptoms. Cognitive and somatic anxiety symptoms as measured by the BAI showed improvements in both CGT and IPT. We previously reported marginally better outcome for CGT versus IPT completers on the BAI total score. In the present study, cognitive symptoms of anxiety showed statistically significantly better results for CGT, whereas the somatic subscale of the BAI showed no difference between groups. This study has several limitations. By design, we were interested in focusing on change in symptoms amongst those who received the assigned treatment, so our sample included only completers. Although an analysis of baseline variables showed no difference between completers and dropouts on any measure, we did not have information on change in associated symptoms for those who discontinued treatment, so we do not know if results would be the same for an intent-to-treat sample. Additionally, our sample size of 69 is relatively small, and power to detect differences between the treatment groups is limited. Moreover, two of the instruments we used in our analyses (SCI-CG and GRAQ) were developed after we began the study, and the sample size for the variables created from these scales was somewhat lower (n = 54). This study of anxiety and depression outcomes in treatment completers shows that CGT was more effective than IPT in targeting these associated symptoms, possibly because CGT specifically targets dysfunctional thoughts and avoidance behaviours as manifestations of grief complications. Additionally, IPT has been repeatedly shown to be an efficacious treatment for depression and our results confirm the importance of distinguishing CG from MDD. Clin. Psychol. Psychother. 23, 118–124 (2016)
124
REFERENCES Beck, A., Epstein, N., Brown, G., & Steer, R. (1988). An inventory for measuring clinical anxiety: Psychometric properties. Journal of Clinical Psychiatry, 56(6), 893–897. Beck, A., Rush, A., Shawb,F., & Emery,G. (1979). Cognitive therapy of depression. New York: Guilford. Boelen, P., de Keijser, J., van den Hout, M., & van den Bout, J. (2007). Treatment of complicated grief: A comparison between cognitive–behavioral therapy and supportive counseling. Journal of Consulting and Clinical Psychology, 75(2), 277–284. Boelen, P., de Keijser, J., van den Hout, M., & van den Bout, J. (2011). Factors associated with outcome of cognitive– behavioral therapy for complicated grief: A preliminary study. Clinical Psychology and Psychotherapy, 18, 284–291. Bui, E., Robinaugh, D., Mauro, C., Wang, Y., Zeng, D., Duan, N., Reynolds, C., Zisook, S., Lebowitz, B., Simon, N. & Shear, M.K. (2014). Development and preliminary psychometric properties for a Structured Clinical Interview for Complicated Grief. Paper presented at the 30th Annual Meeting of the International Society for Traumatic Stress Studies (poster presentation), Miami, FL. Hamilton, M. (1960). A rating scale for depression. Journal of Neurology, Neurosurgery and Psychiatry, 25, 56–67. Hewitt, P., & Norton, G. (1993). The Beck Anxiety Inventory: A psychometric analysis. Psychological Assessment, 5(4), 408–412. Kersting, A. (2011). Prevalence of complicated grief in a representative population-based sample. Journal of Affective Disorders, 131, 339–343. Maercker, A., Brewin, C. R., Bryant, R. A., Cloitre, M., Reed, G. M., van Ommeren, M., et al. (2013). Proposals for mental disorders specifically associated with stress in the International
Copyright © 2014 John Wiley & Sons, Ltd.
K. Glickman et al. Classification of Diseases-11. Lancet, 381(9878), 1683–1685. doi: 10.1016/S0140-6736(12)62191-6 Prigerson, H., Bierhals, A., Stanislav, K., Reynolds III, C., Shear, K., Day, N., et al. (1995). Inventory of Complicated Grief: A scale to measure maladaptive symptoms of loss. Psychiatry Research, 59, 65–79. Prigerson, H., Horowitz, M., Jacobs, S., Parks, C., Asian, M., Goodkin, K., et al. (2009). Prolonged grief disorder: Psychometric validation of criteria proposed for DSM-V and ICD-11. PLoS Medicine, 6, 1–12. Shear, K. (2012). Grief and mourning gone awry: Pathway and course of complicated grief. Dialogues in Clinical Neuroscience, 14(2), 119–128. Shear, K., Frank, E., Foa, E., Cherry, E., Reynolds III, C., Vander Bilt, J., et al. (2001). Traumatic grief treatment: A pilot study. American Journal of Psychiatry, 158(9), 1506–1508. Shear, K., Frank, E., Houck, P., & Reynolds III, C. (2005). Treatment of complicated grief: A randomized controlled trial. JAMA, 293(21), 2601–2608. Shear, K., Monk, T., Houck, P., Melhem, N., Frank, E., Reynolds, C., et al. (2007). An attachment-based model of complicated grief including the role of avoidance. European Archives of Psychiatry and Clinical Neuroscience, 257(8), 453–461. Simon, N. M., Shear, K., Fagiolini, A., Frank, E., Zalta, A., Thompson, E., Reynolds III, C., Silowash, R. . (2008). Impact of concurrent naturalistic pharmacotherapy on psychotherapy of complicated grief. Psychiatry Research, 159(1-2), 31. Steer, R. A., Ball, R., & Ranieri, W. (1999). Dimensions of the Beck Depression Inventory-II in clinically depressed outpatients. Journal of Clinical Psychology, 55(1), 117–128. Wagner, B., Knaevelsrud, C., & Maercker, A. (2006). Internetbased cognitive-behavioral therapy for complicated grief: A randomized controlled trial. Death Studies, 30, 429–453.
Clin. Psychol. Psychother. 23, 118–124 (2016)
