Prolonged grief disorder (PGD) is a novel diagnostic entity proposed to describe a psychiatric disorder that occurs after the loss of a loved one. The disorder is to be implemented in The International Classification of Diseases 11th Revision (ICD-11), scheduled for release in 2018. Failure to recognize and treat PGD may have negative effects on health, functional capacity and quality of life. However, very little is known about the characteristics and treatment of PGD in adult Danish populations although such knowledge is much needed before introducing the disorder in Danish health services.

The present study aims to develop, test and disseminate a group-based treatment manual for PGD in bereaved partners and their adult children. People with high levels of symptoms will be recruited from a large-scale survey study examining frequency, predictors and prognosis of PGD symptoms (NCT03049007). Participants will be randomized to group-based compassion-focused therapy (CFT) for grief or a waitlist control group. Treatment effects will be estimated with prolonged grief symptoms as the primary outcome. Secondary outcomes include depressive symptoms, posttraumatic stress symptoms, symptoms of anxiety, and quality-of-life. The investigators expect that CFT will yield statistically significant effects on prolonged grief symptom compared with the waitlist control group. If CFT is shown to be efficacious in reducing prolonged grief symptoms, the investigators will conduct moderation and mediation analyses with the aim of identifying what works, for whom. Finally, the investigators will perform cost-effectiveness analyses by linking the data with healthcare utilization data from the Danish National registries.

Aims

Our primary hypotheses are:

• CFT will have a statistically significant and long-term effect on PGD symptoms compared with the waitlist control group

Secondary hypotheses are:

• Self-compassion self-criticism, rumination, and exposure to the loss will mediate the effects of CFT on PGD symptoms
• Baseline PGD symptom level will moderate the intervention effects of CFT, with higher symptom level predicting larger symptom improvement
• Gender, age, type of loss, attachment style, neuroticism, positive affect and social support will predict who will benefit most from treatment (moderator analyses)
• CFT will reduce levels of depression, anxiety, and posttraumatic stress symptoms and increase quality-of-life (QoL) compared with the wait-list control group
• Time spent on homework practice will moderate the intervention effects of CFT, with more time spent on homework predicting larger symptom improvement
• CFT is a cost-effective psychological treatment compared with the waitlist-control group

Design

The present project will be conducted as a randomized waitlist-controlled trial, with a 1:1 allocation ratio and using block randomization conducted by an external biostatistician.

Participants

The present randomized controlled trial (RCT) recruits patients from a large survey that invites all bereaved partners in the Central Denmark Region and their adult children (age > 18 years) to participate in a survey regarding trajectories of grief. This means that the present study recruits patients from an existing survey (NCT03049007). All bereaved partners and their adult children from the existing survey study with clinically relevant levels of PGD 11 months after their loss (min.score >= 25 on PG-13) will be invited to participate in the present RCT.

Assessment points

Assessment of all three intervention groups will be provided at five points: (1) before the intervention (baseline), (2) after each session (the CFT group only), (3) after the last intervention session (post-intervention), (4) 3 months post-intervention, and (5) 6 months post-intervention. In addition, data on healthcare utilization will be retrieved from the Danish registries concerning health care services (e.g., visit to the general practitioner, use of medication, number of hospitalizations).

Sample size

Previous studies have generally found large effects of psychological interventions for PGD that included cognitive techniques with effect-sizes (Cohens' d) ranging from 0.80-2.41. Although previous studies testing the efficacy of psychological treatments for prolonged grief have reported large effects, we undertake a more conservative estimate as CFT, so far, has not been tested in a bereaved population.

So far, no values for minimal clinically important difference (MCID) have been established for the PG-13. When such indicators are missing, the general literature on interpretation of clinical change suggests that a change corresponding to one half standard deviation could be a suitable substitute (Copay et al., 2007; Norman et al. 2003). We therefore aim to be able to detect a clinically relevant effect of Cohen's d 0.50. A final sample of 156 participants (2 x 78, 1:1 allocation) will enable us, in a pre-post repeated-measures design, to detect a difference between CFT and the control group, with a two-sided alpha of 0.05, a pre-post correlation (Rho) of 0.5, a statistical power of 0.80, and an estimated dropout rate of 20%.

Statistical analyses

Analyses will be conducted using the statistical software IBM SPSS statistics, v.21 (IMB, Chicago, IL) and Stata v.13 (StataCorp, College Stadium, TX). Baseline group differences will be explored with t tests and χ²-tests. All main effects will be analyzed using Mixed Linear Models (MLMs) based on the intent-to-treat sample, comparing groups over time on all outcome variables. Each hypothesis will be tested in separate models.

Intervention effects will be indicated by a statistically significant 2-way group x time interaction, and results will be interpreted in terms of their effect sizes. As we randomize the participants, we do not expect that any differences in use of alternative treatments prior to or during the study period will exist between study groups. However, in case of an imbalance on this variable, we will include it as a covariate in a second set of analyses.

Possible mediating and moderating effects will be explored for statistically significant outcomes in the main analyses, using MLMs. Cost-effectiveness analyses will be investigated by aggregating healthcare utilization costs and by investigating the effect of CFT on Quality-Adjusted Life-Years (QALY).The probability of CFT being cost-effective are investigated by calculating Cost-Effectiveness Acceptability Curves (CEACs).

Inclusion Criteria:

• Residence in the Central Region of Denmark
• Status of partner or child to the deceased
• A min. score of 25 on the PG-13 at 11 months after the loss

Exclusion Criteria:

• Insufficient ability to understand Danish
• Serious psychiatric disease, e.g., schizophrenia, bipolar disease
• Acute suicidal ideation
• Serious cognitive impairment, e.g., Alzheimer's disease
• Active treatment for serious disease, e.g., chemotherapy for cancer