Complicated grief group therapy for community-residing persons diagnosed with serious mental illness

ABSTRACT

The traumatic death of a significant person is an under-appreciated adverse life event. Unresolved grief from traumatic death is associated with complicated grief (CG) in adulthood and both contribute to and sustain serious mental illness in adulthood. Persons diagnosed with serious mental illness experiencing CG represent vulnerable persons for whom traditional care is therapeutically insufficient, when the trauma of underlying grief is unaddressed. We examined the feasibility and impact of community implementation of Complicated Grief Group Therapy for persons diagnosed with serious mental illness with CG. Program completers showed statistically significant improvement in grief measures and interpersonal well-being.

KEYWORDS:

• Grief
• complicated grief
• group therapy
• intervention
• community-based
• trauma

Background

In 2019, there were an estimated 51.5 million adults aged 18 years or older in the United States with any mental, emotional, or behavioral disorder. This number represented 20.6% of all U.S. adults (National Survey on Drug Use and Health; Substance Abuse and Mental Health Services Administration, [SAMHSA] 2019). Persons diagnosed with serious mental illness (SMI), those having enduring, life-impairing mood and thought disorders, represent 5.2% of adults in America. This represents an estimated 13.1 million adults aged 18 years or older in the United States with serious mental illness (SMI). This number represented 5.2% of all U.S. adults. By definition, SMI is mental health disorder that causes serious functional impairment. Prevalence of SMI was higher among females (6.5%) than males (3.9%), and younger adults aged 18–25 years had the highest prevalence of SMI (8.6%) compared to adults aged 26–49 years (6.8%) and aged 50 and older (2.9%) (National Survey on Drug Use and Health; Substance Abuse and Mental Health Services Administration, 2019).

Adverse life experiences are a significant contributing cause of mental illness, and may eclipse genetic predisposition and all other causal factors (Chapman et al., 2004; Kristensen, Weisaeth, & Heir, 2012; Sachs-Ericsson, Rushing, Stanley, & Sheffler, 2016). The death of a close other (e.g., parents, sibling, and close friend), particularly in traumatic death such as suicide, drug overdose, or homicide, is an underappreciated adverse life event. Unresolved grief from traumatic death is highly associated with complicated grief in adulthood and both contribute to and sustain serious mental illness in adulthood (Sung et al., 2011). Complicated grief (CG) is a debilitating bereavement experience resulting in functional and social impairment. Unlike normal grief, CG is characterized by maladaptive thoughts, feelings, and behaviors that obstruct the difficult, but natural, progression of grief (Parker, McCraw, & Paterson, 2015; Parker, Paterson, & Hadzi-Pavlovic, 2015). Prevalence estimates of CG range from 7% to 20% (Kersting, Brahler, Glaesmer, & Wagner, 2011) in community well populations. Epidemiological data, notably obtained prior to more the recent diagnostic clarity provided in ICD 11 (World Health Organization, 2019), indicated that persons with prior diagnoses of major depressive disorder had higher rates of threshold CG (25.0% vs. 2.8%) and increased prevalence of lifetime alcohol dependence, greater exposure to traumatic events, and lower perceived social support (Sung et al., 2011).

While awareness of the relationship between adverse life events and mental illness is steadily increasing, clinical and research efforts have focused on early identification and intervention in children. Many adults diagnosed with serious mental illness continue to suffer in sorrow, with the symptoms of their mental illness treated, with little attention to their underlying problematic grief (Fisher et al., 2020; Kristensen et al., 2012; Nakajima, Ito, Shirai, & Konishi, 2012).

Persons diagnosed with serious mental illness experiencing CG represent a large, vulnerable population of persons for whom traditional care is therapeutically insufficient (Kersting et al., 2011), resulting in an underserved population with unmet needs (Shear, Muldberg, & Periyakoil, 2017). Given the significantly higher risk of suicide, as much as a threefold increase among persons with traumatic death exposure (Afzali, Sunderland, Batterham, Carragher, & Slade, 2017; Jordan, 2008; Park, Hong, Jeon, Seong, & Cho, 2015; Runeson & Asberg, 2003), effective therapy may contribute to a comprehensive approach to mental health and suicide prevention (Dias De Mattos Souza, Lopez Molina, Azevedo Da Silva, & Jansen, 2016; Mason, Tofthagen, & Buck, 2020).

This project stands out from conventional models and strategies, because no evidence-based group psychotherapy intervention for CG is currently community dissemination-ready to address this problem. Three validated individual psychotherapeutic approaches for treating complicated grief have been reported in the research literature: Complicated Grief Therapy Cognitive Behavior Therapy for Grief, and Meaning Reconstruction Therapy. These three psychotherapies and their distinct and convergent treatment elements have been summarized by the clinical researchers who developed (Shear, Boelen, & Neimeyer, 2011).

Targeted group psychotherapy treatment of CG may prove effective in mitigating the effects of the adverse life experience of traumatic death exposure. Group interventions bring the additional benefits of shared experience, reduction of social isolation, and cost savings to the care of persons with CG. The purpose of this project was to implement our efficacious complicated grief group therapy (CGGT) for persons diagnosed with serious mental illness receiving community-based services and to evaluate feasibility and clinical outcomes of CGGT in this population.

Prior research

We developed CGGT as a group-therapy adaptation of Shear’s complicated grief therapy for individual psychotherapy, the only treatment for CG with established effectiveness (Shear, Frank, Houck, & Reynolds, 2005; Shear et al., 2016). This study builds on our prior lab-based investigations that demonstrated the efficacy of CGGT with clinically significant improvement in CG with bereaved older adults (Supiano & Luptak, 2014), with bereaved dementia family caregivers (Supiano, Andersen, & Haynes, 2015; Supiano, Haynes & Pond, 2017); and with suicide loss survivors (Supiano, Haynes, & Larsen, 2017a; Supiano, Haynes, & Pond, 2017b) (Table 1). These studies with distinct populations, older adults, suicide loss survivors, and bereaved dementia family caregivers represent our foundational work in the highly iterative process of behavioral-intervention implementation. Clinical change in participants confirmed significant reduction in CG on the Inventory of Complicated Grief-revised (ICG-r; Prigerson et al., 1995). In addition to analyzing the outcome of CGGT, we analyzed the process of therapeutic change in CGGT in the suicide-survivor study and the bereaved-dementia-caregiver study, dismantling treatment effects and determining the nature of clinical change at key inflection points of therapy.

Table 1. Summary of efficacy results from three pilot complicated grief group therapy studies

Study Population	N			BGQ						PG-13
					Control Mean (sd)		Treatment Mean (sd)	Effect Size	Control Mean (sd)			Treatment Mean (sd)	Effect Size
Older Adults	34		Pretest Posttest		7.29 (1.40) 6.53 (1.37)		7.06 (1.29) 3.3 (1.30)*	d = 2.42	38.79 (9.46) 30.29 (9.37)			41.5 (5.42) 23.00 (9.02)*	d = 1.34
						BGQ			ICG-r
Suicide Survivors	46		Pretest Posttest		6.73 (2.10) 7.29 (2.78)		7.29(1.90) 4.03 (2.17)*	d = 1.376	42.93 (12.82) 44.52 (16.48)		44.52 (10.17) 28.55 (11.73)*		d = 1.306
Dementia Caregivers	37		Pretest Posttest		5.58 (0.38) 6.29 (0.49)		6.29 (0.32) 3.00 (0.37)	d = 1.324	28.48 (2.10) 31.73 (2.67)		31.73 (1.39) 19.21 (1.60)*		d = 1.351

* indicates p-value for difference between treatment group pre and end of treatment time points is < 0.001.

BGQ = Brief Grief Questionnaire; PG-13 = Prolonged Grief Disorder Scale (used instead of ICG-r); ICG-r = Inventory of Complicated Grief-revised.

Theoretical framework and application to community practice

In developing CGGT, we synthesized two theoretical frameworks that conceptualize both normal change in grief and the disruption that characterizes complicated grief. Current understanding of CG is informed by the Dual-Process Theory (Stroebe & Schut, 1999, 2001). This well-substantiated model of grief adjustment consists of three components: loss-orientation coping (LO), restoration-orientation coping (RO), and oscillation between the two. LO coping includes thoughts and emotions about the lost relationship, while RO coping encompasses the adjustments that the griever needs to face in the life without the deceased. Oscillation is the interaction between LO and RO coping whereby the bereaved confronts the loss, alternating with periods of setting aside thoughts of the loss but facing the changed life (Shear, 2010a, 2010b). Risk factors of complicated grief contribute to ineffective LO or RO coping. We suggest that CGGT restored a normative grief process through carefully designed intervention elements that facilitated LO and RO coping and reduced the influence of risk factors.

The second theory of grief, Meaning Reconstruction Theory, was postulated by Neimeyer and associates (Neimeyer, 2000, 2001a.; Park, 2010), which contributes an explanatory meaning-making model of mourning. Meaning-making refers to the capability of grievers to accept the loss, realize growth, and reorganize personal identity in the context of loss (Hibberd, 2013). Considerable research has supported meaning making as a mediator in both CG and constructive grief (Burke et al., 2017; Burke et al., 2015; Coleman & Neimeyer, 2010). Our prior research on the process of CGGT incorporated Meaning Reconstruction Theory to discern the engagement of the LO and RO coping. We investigated the progression of therapeutic change in CGGT participants (Supiano, Haynes, & Pond, 2017a, 2017b) using the Meaning of Loss Codebook (Gillies, Neimeyer, & Milman, 2014), a coding system that characterizes the meanings grievers ascribed to the death. We evaluated coded video segments of group sessions to articulate how intervention elements transformed the grief of participants.

Intervention

The CGGT intervention used in our lab-based RCTs is a group therapy adaptation of Shear’s textbook, Complicated Grief Therapy for individual grief therapy (Shear, 2003). CGGT is 12 weeks in duration with 120-minute sessions. Intervention elements focus on the relationship between caregiver and care recipient, how memories of life together and illness are interpreted, and strategies for creating a life without the person who died. CGGT treatment elements include psychoeducation, motivational interviewing, cognitive-behavioral techniques, prolonged-exposure techniques, memory work, and meaning-reconstruction activities. In addition, participants invite a supportive other to attend two of the group sessions. Each intervention element is designed to facilitate LO coping, RO coping, or a combination of the two. Intervention activities in later sessions are designed to optimize meaning making of the relationship and the integration of memories of the person who died. (Table 2)

Table 2. Outline of CGGT Intervention Sessions

Session 1	Session 7
Psychoeducation Death story #1 & group feedback Explanation homework, distress scoring	Supportive Other visit #2 (half of session) Difficult-memories worksheet & pictures Grief & self-care goals for next session
Session 2	Session 8
Mindfulness education & practice Developing grief goals & self-care goals	Imaginal conversation for half of group & group feedback Working with emotions/thoughts/body sensations Grief & self-care goals for next session
Session 3	Session 9
Supportive Other visit #1 (10 minutes in group) Working with emotions/thoughts/body sensations Integrated grief check Grief & self-care goals for next session	Imaginal conversation for 2nd half of group & group feedback Working with emotions/thoughts/body sensations Grief & self-care goals for next session
Session 4	Session 10
Death story #2 & group feedback Working with emotions/thoughts/body sensations Grief & self-care goals for next session	Death story #4 & group feedback Education on integrated memory Grief & self-care goals for next session
Session 5	Session 11
Pleasant-memory worksheet & pictures Working with emotions/thoughts/body sensations Grief & self-care goals for next session	Integrated-memories worksheet & pictures Grief & self-care goals for next session
Session 6	Session 12
Death story #3 & group feedback Working with emotions/thoughts/body sensations Grief & self-care goals for next session	Bringing the memory of the deceased forward in one’s life Participant accomplishments Grief & self-care goals for the future

Study aims

Given the very long delay in developing, evaluating, implementing, and disseminating clinical mental health interventions, research funders are recommending that both feasibility goals and clinical outcomes are evaluated using the same sample of individuals with the identified condition (Landes, McBain, & Curran, 2019). This preserves the number of persons being studied, particularly important with prevalent but difficult to recruit research participants, expedites the identification of a clinical signal of therapeutic change, and speeds the progress toward hybrid implementation-effectiveness trials. This study examined feasibility and impact of community implementation of CGGT for persons diagnosed with serious mental illness who have experienced a traumatic death and meet clinical criteria for CG. We sought to determine whether CGGT, which has demonstrated efficacy under controlled laboratory conditions, was feasible and had a beneficial effect when implemented in a pilot pragmatic study. We conducted two 12-week CGGT groups.

Specific Aim 1: Conduct feasibility analysis of community delivered CGGT to guide evaluation, refinement, and dissemination. We evaluated recruitment, retention, acceptability, tolerability, safety, clinician trainability, and stakeholder satisfaction.

Specific Aim 2: Evaluate the impact of community delivered CGGT on clinical outcomes. The primary clinical outcome was treatment response, measured as change in scores on the Inventory of Complicated Grief-revised (ICG-r; Prigerson et al., 1995), the Brief Grief Questionnaire (BGQ; Shear & Essock, 2002), and the weekly facilitator generated Clinical Global Impressions Scale (CGI, Guy, 1976). We also analyzed change in depression, anxiety, and suicidality.

Specific Aim 3: Inform further refinement of CGGT and modification of the treatment manual for community use, future translation, and dissemination. We are making modifications to the facilitator and participant CGGT manuals for this population and preparing for broad dissemination to community agencies.

This study was designed and conducted as a Stage IIIa feasibility trial, as defined by the National Institutes of Health Stage Model for Behavioral Intervention Development (Onken, Carroll, Shoham, Cuthbert, & Riddle, 2014).

Method

Sample

CGGT group participants were adult patients of Impact Clinic (formerly Polizzi Clinic), Utah’s only outpatient psychiatric clinic providing stabilizing mental health care for low-income, uninsured persons with untreated serious mental illness. The client population of Impact Clinic includes persons with traumatic social histories, minimal social supports, high levels of substance use disorders, homelessness, and chronic medical illnesses in addition to long-standing mental illness. Potential participants had decisional capacity, as they are agreeing/consenting to receiving their usual psychiatric care as determined by the medical director (psychiatrist) at Impact Clinic; therefore, persons were not cognitively impaired/vulnerable. As identified by clinical staff of Impact Clinic, potential participants had 1) experienced a traumatic life death and 2) met clinical criteria for complicated grief (score of 5 or greater on Brief Grief Questionnaire and score of 24 or greater on Inventory of Complicated Grief-revised) as assessed by the Research Assistant (RA). All participants were reported by Impact Clinic providers as diagnosed with serious mental illness, either serious mood or thought disorder.

Supportive other participants: One treatment element of CGGT requires the inclusion of a supportive family member or friend for each participant into the group on two occasions. Supportive others will be adult family members or friends selected by group participants, if available. Supportive other participants have no exclusion criteria.

Partnering organizations

Impact Clinic: The community CGGT groups were conducted at the Impact Clinic, which is centrally located with optimal access to mass transit and operates with interdisciplinary staff including a psychiatrist, Psychiatric Advance Practice Nurse faculty, psychiatry residents, psychiatric DNP students, social work students, and undergraduate volunteers. The clinic has participated in prior research studies and serves as an interdisciplinary community clinical training site.

Community-Based Participatory Research (CBPR) Advisory Board: As an important component of our research program, we assembled an active CBPR advisory board in 2013 including representatives of the Utah chapters of the American Foundation for Suicide Prevention and National Alliance on Mental Illness, University mental health services faculty, pilot study facilitators, and former suicide study participants and their selected supportive others. These community partners are committed to assisting with recruitment and contribute expertise in developing culturally sensitive approaches to the intervention, recruitment and retention, community access, and translation and dissemination. No potential nor final participants were known to Advisory Board members.

Study measures

Participants were administered the same assessment instruments at initial enrollment, midpoint (following 6^th session), upon completion of the 12 week group, and at 6-week follow-up. Assessment instruments included the demographic information intake form, the BGQ and ICG-r, the Patient Health Questionnaire-2 (depression, PHQ-2; Kroenke, Spitzer, & Williams, 2003), the Generalized Anxiety Disorder −2 scale (anxiety, GAD-2; Kroenke, Spitzer, Williams, Monahan, & Löwe, 2007), the Interpersonal Needs Questionnaire-short form (suicidality, INQ-sf; Parkhurst, Conwell, & Van Orden, 2016), the Integration of Stressful Life Events-short form (grief stress, ISLES; Holland, 2016), the Suicidal Behaviors Questionnaire-revised (suicidality, SBQ-r; Osman et al., 2001), and the Depressive Symptom Inventory Suicidality Subscale (suicide symptoms, DSI-SS; Von Glischinski, Teismann, Prinz, Gebauer, & Hirschfeld, 2016). Participant progress was assessed weekly by CGGT co-facilitators using the Clinical Global Impression Scale (CGI, Guy, 1976).

Procedure

Recruitment: Recruitment flyers were posted in the clinic and in exam rooms, directing potential participants to discuss the study with their provider and/or contact the PI. Eligibility (stage 1). Potential participants were identified by clinic leadership (Medical Director) as having history of a traumatic life event/death of close other and were diagnosed with serious mental illness, mood or thought disorder, or both. Potential participants received an invitation phone call from the RA. Eligibility (stage 2). Interested individuals were invited to a pre-screening interview with the RA in person at the clinic. Those who met the final inclusion criteria, a minimum score of 5 on BGQ, and a minimum score of 24 on the ICG-r were invited to participate. The RA carefully reviewed the informed consent document. Those who signed the informed consent document were provided with an additional copy of the document and completed the full initial assessment as administered by the RA. Participation was voluntary. There was no delay between informing and obtaining consent. Supportive others were selected by enrolled participants, and their contact information was provided to the study staff. Interested potential supportive other participants were mailed a copy of the informed consent document and returned the completed form at their first CGGT session.

Conducting the CGGT intervention: The two 12-week group cohorts were held in the private conference room of Impact Clinic, a room typically used for staff meetings. All study materials were provided. The two group cohorts received the CGGT intervention February-May 2019 and August-November 2019, respectively.

Compensation: Gift cards were provided to participants according to a compensation schedule that included weekly 5 USD gas cards, and 10 USD gift cards at each testing interval. Participants who withdrew from the CGGT group, but continued the assessment process, were compensated on this schedule.

Treatment Fidelity: To assure facilitator competence, adherence, and quality of implementation, the facilitators had weekly peer supervision sessions. In addition, two randomly selected sessions of each 12-week group (four total over the study) were audiotaped and evaluated for treatment fidelity using an audit and feedback tool (Mignogna et al., 2014) by an independent evaluator familiar with CG therapies, but unaffiliated with the study.

Project team: Our study team included the Principal Investigator, a nurse scientist co-investigator, and a PMH-DNP student who served as Research Assistant. Both CGGT groups were facilitated by a highly skilled LCSW, who had conducted several CGGT groups in earlier research studies (older adults and survivors of suicide loss). Two PMH-DNP students with advanced standing served as CGGT co-facilitators.

This study was approved by the University of Utah Institutional Review Board, IRB_00116449.

Results

Establishing the relationship with impact clinic

Impact Clinic is a not-for-profit program and is funded almost entirely by donations. It is staffed by one part-time semi-retired psychiatrist, who has adjunct faculty status at the University of Utah School of Medicine, one part-time administrator, one AmeriCorps Vista volunteer program director, and a part-time receptionist. At the time of our project, the clinic served as a clinical rotation site for psychiatric residents and PMH-DNP students. The educational structure of the clinic provides trainees with clinical practice in comprehensive mental health assessment and prescriptive practice. The emphasis of the clinic was and remains immediate psychiatric assessment and medication management. Little or no psychotherapies are provided. Little or no case management services are provided. Once clients obtain insurance, they leave the clinic, as they no longer qualify for care, but most leave without notifying the clinic. No show rates for scheduled clinic appointments were very high, and despite instituting a by-phone appointment reminder system halfway through our project, no show rates remained unchanged. As is common in this treatment population, many clients do not have phones or cannot afford to keep their phones and phone numbers change frequently or are disconnected. The leadership and staff of Impact Clinic were supportive throughout the study and appreciative of our efforts to provide CGGT. The medical director and the PI met nearly every week to review goals and progress of both clients and the project.

Feasibility

Recruitment and retention: Impact Clinic provided the study population we were seeking, as the purpose of a feasibility study is to target the population of interest, and for us, this population was persons with complex and serious mental health conditions, persons who had multiple, complex traumatic losses, and persons who were socially and economically under-resourced. With assistance of Impact Clinic staff, we were able to identify 188 clients likely to qualify for the study. This population presented numerous challenges for recruitment. Most potential participants were experiencing psychological instability, most were adjusting to recent psychiatric hospital discharge, new medications, and new providers, and continued to have the same life circumstances that precipitated and maintained poorer function (unemployment, few social supports, poor health, and financial instability). The average number of lifespan traumatic losses identified in our sample was four, which is among persons with an average age of 38 years. These losses included deaths by suicide and homicide, as well as physical and psychological abuse, personal and familial substance abuse, parental abandonment, divorce, job loss, and multiple medical comorbidities. As mentioned earlier, phone and e-mail access remained problematic. Many potential participants did not answer phone or text calls, likely fearing dunning calls from collection agents. While we did not have nor request access to psychiatric medical records, all potential participants qualified for clinic care on the basis of diagnosis of serious mental health conditions. Many had co-morbid substance use disorders. Many had extensive history of suicidality and prior suicide attempts. The Impact Clinic population represents a very common client base of community mental health clinics across the United States.

Our RA made concerted efforts at engagement and recruitment of potential participants. Per-clinic protocol, phone contacts and interviews could only be initiated within the clinic setting, necessitating numerous trips to the clinic. Potential participants were difficult to contact, difficult to schedule, and frequently no-showed for scheduled enrollment interviews (see Table 3).

Table 3. Recruitment

Category	N	Notes
Total potential participants	188 (%)
Not interested	12 (6.4%)	Was directly told that they weren’t interested
Contacted by lost to follow up after initial contact	33 (17.6%)	Would ask to call back at a different time, would call but then not answer; would tell me they would call back, but wouldn’t – I would call back and no answer
Could never contact	75 (39.9%)	Left VM’s, busy signal, mailbox full, sent e-mails, etc. never had any kind of response
Number not in service, clinic didn’t have a phone number, or wrong number	16 (8.5%)
Scheduled (came)	18 (9.6%)	1 was not eligible 1 had school and couldn’t make the group time
Scheduled (never came)	14 (7.4%)	They would sometimes keep rescheduling but never come, scheduled once, no-showed, and wouldn’t answer follow up calls.
Work/school conflicted with group time	9 (4.8%)
No longer with Polizzi Clinic	2 (1.1%)
Moved	4 (2.1%)	Some of these were people who gave the number of a residential home or treatment center they were staying at and had since left the facility
Didn’t speak English	3 (1.6%)
Too young	2 (1.1%)

When interviews occurred, our RA was able to establish rapport and complete the assessment instruments. These potential participants voiced appreciation, endorsed the trauma in the life histories, and expressed a desire for obtaining psychotherapeutic support. Of the 188 potential participants, 17 were fully assessed for participation and 16 were eligible, consented, and enrolled (see Table 4). Of these, 10 persons participated in the two CGGT groups.

Table 4. Participant enrollment and retention

Category	N	Notes
Individuals eligible	16/17	All but 1 who were assessed were eligible
Never showed up	3	We were able to continue to get data from 2, but never heard from a third after she said she couldn’t make the first session.
Lost to follow up	3	No phone or e-mail, no response to contact by text or call
Completers	10

Acceptability, Tolerability, and Safety: The intervention, CGGT was regarded as welcome, “overdue,” necessary, and was simultaneously deeply challenging for all participants. Avoidant behavior is a key feature of complicated grief and in psychological trauma that is refractory to treatment. Each participant endorsed a “push-pull’ phenomenon prior to and within each session – wanting to join the group, wanting to get and give support, wanting to face and resolve their losses, while concurrently wanting it to “be over,” to “just get past it,” or “to never have to think about it.” Participants were strongly urged to participate in all treatment elements, but were given exceptional latitude by facilitators to participate “as much as you can,” an approach much less necessary in our lab-based studies, but congruent with the clinical needs of study participants. Facilitators were deeply invested, encouraging, flexible, and kind. Participants reported feeling affirmed by facilitators and connected with each other, occasionally checking in on each other and sharing rides to clinic. All participants valued the group and the intervention, as well as the facilitators and each other. Participants valued the gift cards (for attendance and to fund gasoline) and the snacks provided. We concluded that providing a full meal at each session would have been an appropriate and helpful addition to our protocol.

All participants were routinely assessed for suicidality. We had no expressed suicidal intent nor attempts during the treatment course or upon follow-up in group participants.

Sample

The demographics of our sample of study participants are presented in Table 5. Our sample had a mean age (SD) of 34.76 (11.95) years and age range of 19–58 years. Our participants were mostly female (N = 11, 68.8%) and white (N = 12, 75%). Notably, while all had multiple losses, the loss of chief importance identified was a person who was the primary source of emotional and material support. Of reported deaths, 87% were described as both unexpected and traumatic. Half of the participants said that their health was fair or poor compared to others. The two groups represented 9 and 7 participants, respectively. There were no discernible demographic differences between the two groups, and reported losses and life experiences were similar. CGGT completers were 6 and 4 participants, respectively.

Table 5. Demographics and participant characteristics

Variables		N	%
Age in years
N (Nmiss)		16 (0)
Mean±SD			34.75 ± 11.95
Min-Max			[19– 58]
Gender
Female		11	68.8
Male		3	18.8
Other/Non-binary		2	12.5
Race
White		12	75
Black		1	6.3
Hispanic		3	18.8
Relationship to primary decedent
Sibling		1	6.25
Spouse		2	12.5
Parent		6	37.5
Child		1	6.25
Grandparent		3	18.75
Other		3	18.75
Death unexpected
Yes		14	87.5
No		2	12.5
Family living at home
Yes		11	68.8
No		5	31.3
Health rate compared to others
Excellent		1	6.25
Very good		2	12.5
Good		5	31.25
Fair		3	18.75
Poor		5	31.25

Clinical outcomes

Intervention participation for the 12 CGGT sessions attended varied from 0 to 12 with M = 6, SD = 4.21, and Mdn = 7 sessions. Importantly, 3 individuals did not attend any sessions, 5 attended 1–6 sessions, and 8 attended 8 or more sessions.

Participants were examined for changes in several key behavioral questionnaires prior to the intervention (n =16) and at the end of the study (n =10). Paired Wilcoxon signed rank tests were utilized to examine changes in scores on the Brief Grief Questionnaire (BGQ), Inventory of Complicated Grief (ICG-r), Patient Health Questionnaire (PHQ-2), Generalized Anxiety Disorder scale (GAD-2), Integration of Stressful Life Experiences scale (ISLES), and and Interpersonal Needs Questionnaire (INQ). Table 6 displays the median values, Z statistics, and p statistics. Statistically significant improvement at the end of the timeline compared to baseline at p <.05 was observed for BQG, ICG, and INQ questionnaires. ISLES, PHQ-2, and GAD2 were not statistically significant.

Table 6. Descriptives and Wilcoxon Sign rank tests for questionnaire outcomes

	BGQ PRE	BGQ POST	ICG PRE	ICG POST	PHQ-2 PRE	PHQ-2 POST	GAD2 PRE	GAD2 POST	ISLES PRE	ISLES POST	INQ PRE	INQ POST
N	16	10	16	10	16	10	16	10	16	10	15	10
Mean	7.38	3.90	38.56	27.80	3.50	2.50	4.00	3.70	15.63	19.70	59.21	49.18
Std. Deviation	1.78	1.85	9.77	8.31	1.97	1.90	1.86	1.83	5.06	3.92	8.74	7.47
Minimum	4	1	24	13	0	0	1	0	7	13	45.00	34.29
Maximum	10	6	55	39	6	6	6	6	25	25	75.00	56.79
Median	7	4.5	38.5	28.5	3	2	4	4	14.5	20	57.86	50.89
N (post-pre)		10		10		10		10		10		10
Z value		−2.687		−1.99		−0.923		−0.566		−1.939		−2.68
p statistic		0.007		0.047		0.356		0.571		0.052		0.01

Following each session, the two facilitators debriefed on each participant’s response to the intervention and rated each participant’s current grief severity and grief improvement using the CCGI scale. Figures 1 and 2 display severity and improvement in the facilitator measured CGI scale over the period of the 12 sessions. Over time, there is trend toward decrease in grief severity, as measured by the CGI. For CGI improvement, 4 represents “no improvement,” and therefore, at each of the sessions, there is a trend for individuals to have continued improvement. Owing to missingness, we could not conduct linear regression to examine change over time. Wilcoxon signed rank test found decrease in CGI severity (n =7) from Session 1 (Mdn = 6) to Session 12 (Mdn = 3), Z =−2.932, and p =.017. For CGI improvement, at session 2 (n = 6), 3 individuals were minimally worse and 3 individuals were minimally improved; at session 6 (n = 6), 1 showed no improvement and 5 were minimally improved; and at session 12 (n =9) 3 were much improved and 6 were very much improved (Figures 1 and 2).

Figure 1. Change in grief severity as measured on clinician global impressions scale

Figure 2. Change in grief improvement as measured on clinician global impressions scale

The participant experience

The impact of the group intervention on participants, even those with less-than-ideal attendance, was readily observable by co-facilitators throughout its duration. While we conducted brief exit interviews with participants, we did not record interviews nor perform formal qualitative analysis of narrative content either during or after the group. The following two briefcase summaries were synthesized as representative of the participant experience by student co-facilitators.

Jane: One especially shy and socially anxious participant, who was grieving the loss of abandonment by her emotionally distant father, initially made very few contributions to group discussions and, when she did, whispered them with timidity. Her poor eye contact and tearful nature communicated the shame she felt for having been “replaced” by her father’s stepdaughter. Her father died two years prior to the group initiation, but she recounted his abandonment as the relevant trauma. Co-facilitators actively invited her participation but she frequently opted instead to avoid emotional flooding through passive observation. Her emotional restriction made it difficult to accurately assess her internal reactions to session content. Still, she demonstrated increased openness, however slight, with each subsequent session. Her voice became increasingly clear, confident, and unashamed. She smiled more often and began sharing jokes, even in the midst of tears. Though she missed session eight due to family obligations, she returned for session nine and surprised co-facilitators and peers alike by engaging in the imaginal conversation with striking vulnerability. By the group’s end, she exhibited a marked acceptance of her loss and enhanced self-worth. She relinquished her need for parental validation and instead found healing by establishing a warm, sisterly relationship with the step-daughter who “replaced” her. This group clearly affected this participant in potentially life-changing ways.

Pamela: lost her mother to suicide in her early twenties and was initially hesitant about being in the group due to the seven years that had passed since the death. As the group went on, she was able to find value in the tasks of the group and realized emotional growth as a result. Pamela was able to move through many of the suppressed feelings surrounding her mother’s death, primarily guilt. She was able to connect with her brother on another level by asking him for the whereabouts of one of their mother’s items. Pamela and her brother were grateful for the ability to revisit memories and address the feelings surrounding their mother’s death. She was highly motivated to complete the associated homework each week and had the highest attendance of group participants, actions that reflect the benefit she was receiving from the group.

Treatment fidelity

In each 12-week CGGT group, two sessions were randomly selected to be audio-recorded and evaluated for facilitator manual adherence and skills performance by the RA and a second evaluator familiar with the CGGT intervention but unaffiliated with the study. Manual adherence and skills performance were assessed using the Manual Adherence and Skill Performance Checklist (Mignogna et al., 2014). Manual adherence was 88%, and Skills Performance was 95%.

Student trainability and educational outcomes

The three graduate students involved in the study gained considerable skills and confidence. All were Collaborative Institutional Training Initiative (CITI) certified to participate in human research. Our graduate research assistant learned research methods including standards for protection of human subjects, IRB procedures, recruitment, consenting and enrollment procedures, clinical research interviewing and assessment skills, and data entry, management, and analysis. Our clinical trainees received a shared clinical leadership opportunity from a master clinician highly skilled in CGGT, as well as exposure to a difficult-to-treat population. In addition to debriefing following each group session, all students participated in a half-day comprehensive retreat to review clinical notes, dismantle all CGGT treatment elements, discuss the impact of each element and procedure, make recommendations to the treatment protocol specific to this population, and prepare content for a manuscript. Students achieved program learning objectives in knowledge, attitudes, and skills.

Discussion

Despite the magnitude of the trauma and difficult life circumstances experienced by our CGGT participants, all completers achieved not only significant improvement across key measures for complicated grief and interpersonal needs but also significant improvement that was observed as symptom reduction and restoration of function in the areas measured. Practically, for our participants, this meant that the grief and difficult processing of their losses continue, but they proceed in a healthy way with skills, confidence, and steady gains toward integrated grief. Participants also made near significant gains in integration of stressful life events, indicating that the skills learned in CGGT address not only the traumatic loss(es) discussed in the group but impact-associated past and present life stressors. CGGT is not intended to treat anxiety or depression, yet it is notable that the potentially triggering content of group discussion, particularly in the presence of avoidance tendencies inherent in complicated grief, did not precipitate relapse into worsened depression, anxiety, or suicidality. In fact, participants realized stability in these areas. Each group cohort achieved a level of cohesiveness, as observed by facilitators and evidenced by the mutual support that is a valuable component of group therapy. While there were enormous challenges in gathering and maintaining both CGGT groups, it is clear that Complicated Grief Group Therapy is both efficacious and clinically appropriate in this high-need, underserved population.

The Impact Clinic was favorably affected by the partnership with our “Complicated Grief Group Therapy for Community-residing Persons Diagnosed with Serious Mental Illness” study. The clinic Board of Directors adopted a goal of hiring a Licensed Clinical Social Worker to provide case management to monitor and track clinic patients and is now an MSW field placement site. As trauma-informed group therapy practices are further integrated into the program, we expect that CGGT will be a therapeutic modality co-facilitated by Impact Clinic clinical trainees in the near future. Impact Clinic recognizes the need for wrap-around care to accomplish its primary purpose in early assessment and medication stabilization, and our partnership will be enduring to our mutual benefit and that of the clinic clients.

We continue to make revisions to the Complicated Grief Group Therapy Manual for facilitators and the manual for participants. Notable revisions include simplification of participant homework, provision to accommodate situations where participants cannot identify a “supportive other” (this was the case for all but 3 study participants), and structuring the possibility of conducting the intervention twice/week over 8 weeks to accommodate shorter time clients who may receive services in the agency. These modifications will yield facilitator and participant CGGT manuals suitable for translation into wider community practice.

Study strengths and limitations

As a pilot pragmatic feasibility study, the statistical analyses were not powered to determine effectiveness of the intervention. We were able to identify and target our population of interest – persons diagnosed with serious mental illness facing multiple life challenges, and meeting clinical criteria for complicated grief. This treatment population is difficult to access, recruit, and retain for a variety of reasons, and our retention rate was 63%. Attrition is frequently a concern for statistical analyses in clinical intervention studies, frequently exceeding 30% of the sample (White et al., 2010). This rate of attrition refers to lab-based investigations, and is normally much higher in community-based trials. Due to our effective partnership with Impact Clinic, several clinic structure and design recommendations may mitigate these recruitment and retention issues, for research and patient care, a worthy goal of feasibility trials. Participants found the therapy challenging, yet voiced satisfaction and a sense of accomplishment. CGGT completers evidenced clear benefits to their grief, their mental health, and their well-being, substantiating the benefit of the intervention. The educational aim was accomplished, and more grief content, group therapy skills, and student practice opportunities to learn CGGT will be incorporated into clinical traineeships, for subsequent dissemination into other programs.

Conclusion

We are pleased to report that this modest implementation of CGGT suggests promise for addressing traumatic death in persons diagnosed with serious mental illness and is suitable for implementation in community mental health practice settings and merits further study and wider dissemination. Notably, concerns of recruitment and treatment completion reflect the particular challenges of the client population understudy; one that is under-resourced (health insurance, access to medical care, transportation, housing) and many includes difficulties with relationships both among the living and the deceased. As a preliminary pragmatic pilot study conducted with a high-need sample of persons diagnosed with serious mental illness, we established acceptable feasibility and obtained a “signal” of preliminary clinical value that supports further investigation of Complicated Grief Group Therapy in a fully powered hybrid-implementation effectiveness trial.

Conflicts of interest

No potential conflict of interest was reported by the author(s).

Additional information

Funding

This project was funded by The Rita And Alex Hillman Foundation, as an “Emergent Innovation Program Grant,” and by the Robert and Carma Kent Award for Suicide Research.