Efficacy of an integrative CBT for prolonged grief disorder: A long-term follow-up

Highlights

• Patients with prolonged grief disorder (PGD) had participated in a randomized controlled trial.

• They were followed up 1.5 years after integrative cognitive behavioral therapy for PGD.

• Post and follow-up ES for prolonged grief symptoms were large.

• The pre to post-improvement in overall mental health was maintained.

• This is the largest PGD treatment sample followed up for this long so far.

Abstract

Background

While some intervention trials have demonstrated efficacy in treating prolonged grief disorder (PGD), data on long-term treatment effects are scarce.

Methods

Fifty-one outpatients with clinically relevant prolonged grief symptoms, who had participated in a randomized controlled trial (RCT), were followed up, on average, 1.5 years after integrative cognitive behavioral therapy for PGD (PG-CBT). Initial assessment procedures were repeated, with PGD symptom severity as the main outcome and general mental health symptoms as secondary outcomes. As results in the immediate and delayed treatment groups (former wait list) were similar, the follow-up data were pooled.

Results

Overall, 80% of the original ITT sample could be reached, that is 89% of the 37 treated participants, as well as 8 out of 14 participants who had dropped out of the RCT. The considerable short-term treatment success of PG-CBT was stable; pre to follow-up Cohen׳s d was large, with 1.24 in the ITT analysis and 2.22 for completers. The pre to post-improvement in overall mental health was maintained.

Limitations

Since the RCT wait list group had been treated after their waiting period as well, no controlled long-term outcomes are available.

Conclusions

PG-CBT proved to be effective in the longer run. In comparison to other RCTs on prolonged grief this is the largest sample followed up for this long.

Introduction

Over the last two decades, prolonged or complicated grief has emerged as a well-defined mental disorder, distinguishable from major depression and posttraumatic stress disorder or other stress-related or anxiety disorders (for overviews see Bryant, 2014, Jordan and Litz, 2014). Core symptoms are intense yearning and preoccupation with the deceased; reactive distress symptoms, such as feeling stunned or shocked by the loss; avoidance of reminders of the reality of the loss and emotional numbing, and finally social/identity disruption, for instance feeling detached or having difficulties trusting others (Prigerson et al., 2009). Such symptoms should cause clinically significant impairment and persist longer than what would be deemed normative regarding an individual׳s cultural background (Maercker et al., 2013). In ICD-11, where the diagnosis will be included as “prolonged grief disorder” (PGD), the duration criterion will probably be six months after the death of a significant other (Jordan and Litz, 2014). Prevalence rates of prolonged grief were found to be 3.7% in a German sample (Kersting et al., 2011) and 4.8% in a Dutch sample of older adults (Newson et al., 2011). In the case of bereavement, prevalences ranged from 6.7% (Kersting et al., 2011) up to 25.4% in older adults (Newson et al., 2011). PGD has been found to be associated with deteriorated health (Stroebe et al., 2007), increased depression, and suicidality (Boelen and Prigerson, 2007, Latham and Prigerson, 2004).

Meta-analyses on grief interventions indicate that the distinction between normal, non-pathological grief and clinically impairing prolonged grief is important. Whereas non-selective interventions for the bereaved were found to be ineffective at best, psychotherapy for clinically relevant prolonged grief symptoms was found to be at least of moderate effectiveness (Currier et al., 2008; cf. Rosner and Hagl, 2007). The most recent meta-analysis found a non-significant effect size (ES) of 0.03 for preventive studies and a significant, albeit heterogeneous ES of 0.53 for the treatment of clinically relevant prolonged grief (Wittouck et al., 2011). Treatment studies are still rare (in Wittouck et al., 2011), analysis was based on five studies, two of them group interventions). However, there is growing evidence from randomized controlled trials (RCTs) that indicated and specific grief interventions are effective for PGD in different settings, that is, both in individual treatment (e.g., Boelen et al., 2007, Shear et al., 2005) and in group therapy (Supiano and Luptak, 2014), or when delivered via the Internet (Wagner et al., 2006). In newer studies with more recently bereaved individuals, Internet-based grief interventions even showed effectiveness with high risk populations that were not selected for PGD at all (Kersting et al., 2013), or, in terms of indicated prevention, with individuals manifesting grief symptoms on substantial but subthreshold levels (Litz et al., 2014). And recently, a small trial suggested that non-specific behavioral activation (together with functional assessment and education regarding grief symptoms) might suffice to profoundly improve PGD (Papa et al., 2013). Overall, these newer RCTs attest to the efficacy of structured, behavioral interventions for PGD, with moderate to high ES when compared to wait list.

However, studies with long-term follow-ups are rare. In the study by Boelen et al. (2007), a combination of exposure followed by cognitive restructuring resulted in significant improvements in prolonged grief with high pre- to post-treatment ES. These improvements could be maintained up to the 6-month follow-up: Pre- to follow-up assessment Cohen׳s d was 1.82 for completers (n=16) and 1.25 in intention-to-treat (ITT) analysis (n=23). A vice versa combination with cognitive restructuring followed by exposure still achieved ES of 1.41 for completers (n=16) and 0.83 for the ITT sample (n=20). In terms of general psychopathology, the long-term ES in the more successful therapeutic combination were 1.55 for completers and 1.20 for the ITT sample. Wagner and Maercker (2007) conducted a 1.5 year follow-up for their Internet-based grief intervention (Wagner et al., 2006) and reported that all treatment gains could be maintained (completer analysis only; n=22). Finally, Papa et al. (2013) found that the considerable treatment gains in their study could be maintained up to three months (n=20).

In order to develop our intervention for PGD we had reviewed earlier successful trials, together with the results of meta-analyses, to identify the most promising treatment strategies (Rosner and Hagl, 2007). Those were (1) education about normal and prolonged grief processes, (2) exposure to the most painful aspects of the loss, together with cognitive restructuring, and (3) transformation of the loss to enable change. Elements from Gestalt therapy, systemic family therapy and psychodrama were included; so this therapy should be viewed as an integrative cognitive behavioral therapy targeting maladaptive prolonged grief (PG-CBT; Rosner et al., 2011).

PG-CBT was then evaluated in a first RCT, with 51 participants suffering from clinically relevant PGD symptoms (Rosner et al., 2014). Immediate treatment effects were considerable, with within-group effect sizes for the 19 PG-CBT completers of d=1.65 for grief severity (d=1.26, for the ITT sample), and d=0.68 for general mental health symptoms (d=0.64 for ITT). For grief severity, PG-CBT showed a significantly better outcome than the untreated wait list control, with large between-group effect sizes (d=1.61 for completers; d=1.32 in the ITT analysis). For overall general mental health symptoms, PG-CBT did not achieve better outcomes than wait list; however, a look at subscales revealed significantly more improvement in depressive symptoms (d=0.60 for completers; d=0.73 in the ITT analysis). Overall, PG-CBT was shown to be effective and specific for PGD treatment. Below, we report the results of the subsequently treated wait list group of the original RCT, together with long-term follow-up data, pooled for all treated patients. Our research examined whether the initial symptom improvement could be maintained in the long-term.