Introduction
Family caregivers provide indispensable medical and emotional support to people with advanced cancer, often while confronting anticipatory loss and major role changes. Caregiving typically combines complex symptom management, decision support, coordination of services, and sustained emotional labor, which together heighten vulnerability to cumulative stress (Kustanti et al. Reference Kustanti, Fang and Linda Kang2021). Across settings, family caregivers report substantial psychological burden that extends into bereavement, with frequent elevations in anxiety, depression, grief-related symptoms, and guilt (El-Jawahri et al. Reference El-Jawahri, Greer and Park2020). This distress does not cease with the patient’s death but frequently evolves into complicated or prolonged grief disorder, a condition characterized by intense, persistent yearning and functional impairment (Pan and Lin Reference Pan and Lin2022). Prevalent risk factors include prior mental health difficulties, high caregiving demands, and strained family communication, highlighting the need for preventive and early post-loss support within palliative services.
A growing intervention literature targets these risks among family caregivers. Randomized trials of structured, manualized family therapies initiated during hospice care and continued after the death have shown reductions in the severity of complicated grief relative to usual care (Harrop et al. Reference Harrop, Morgan and Longo2020). Additional evidence supports targeted approaches such as family-based dignity interventions that mitigate anticipatory grief in family caregivers of dying patients and therapist-assisted web-based cognitive-behavioral programs that lessen prolonged-grief symptoms after cancer bereavement (Kaiser et al. Reference Kaiser, Nagl and Hoffmann2022; Ghezeljeh et al. Reference Ghezeljeh, Seyedfatemi and Bolhari2023). A later randomized controlled trial (RCT) among families at risk for dysfunctional interaction patterns showed that 10 sessions of the same manualized intervention, begun pre-loss and extended post-loss, lowered the incidence of prolonged grief disorder at 13 months after the patient’s death (Kissane et al. Reference Kissane, Zaider and Li2016). A recent systematic review nonetheless emphasizes variability in treatment outcomes across different grief intervention studies (e.g., complicated grief severity, prolonged grief incidence, and depressive symptoms) and limited standardization of intervention components, suggesting the need for rigorously specified, theory-grounded protocols that bridge anticipatory and early post-loss periods and that focus on family caregivers’ grief responses (Becqué et al. Reference Becqué, van der Wel and Aktan-arslan2023). Cognitive-behavioral therapy (CBT) offers a coherent foundation for such standardization because it targets mechanisms implicated in complicated grief, including maladaptive avoidance, dysfunctional appraisals, negative global beliefs, and difficulties integrating the reality of the loss (Lechner-Meichsner et al. Reference Lechner-Meichsner, Mauro and Skritskaya2021). CBT is therefore particularly well suited to address the mechanisms that maintain distress in bereaved family caregivers. Cognitive-behavioral models of complicated and prolonged grief describe 3 interrelated maintaining processes: incomplete integration of the loss into autobiographical memory, global negative beliefs about the self, world, and future, and anxious or depressive avoidance of painful reminders. These processes map closely onto family caregivers’ experiences of persistent self-blame for perceived caregiving failures, catastrophic interpretations of the patient’s suffering or death, and withdrawal from previously valued activities. Structured CBT interventions explicitly target these mechanisms through psychoeducation, cognitive restructuring of guilt-laden or hopeless appraisals, graded exposure to avoided situations and memories, and behavioral activation techniques that foster reengagement with meaningful roles. In line with these theoretical accounts, randomized trials demonstrate that CBT-based treatments yield greater reductions in complicated grief and associated psychopathology than supportive counseling or nonspecific psychotherapy, and that grief-focused CBT protocols achieve medium-to-large effects on prolonged grief symptoms in meta-analytic syntheses (Buck et al. Reference Buck, Cairns and Emechebe2020; Komischke-Konnerup et al. Reference Komischke-Konnerup, Zachariae and Boelen2024).
Against this background, we designed an RCT conducted from October 2024 to November 2025 that enrolled family caregivers of patients receiving hospice care, randomizing caregivers to an intervention group receiving a structured grief counseling program grounded in CBT or to a control group receiving usual support alone. We hypothesized that, compared with a control group receiving usual support, caregivers in the intervention group would report significantly lower levels of grief responses, significantly lower levels of psychological distress, and significantly higher levels of positive adaptation and health. This study seeks to establish rigorous evidence for a manualized, phase-based intervention that supports caregivers through the critical transition from anticipatory grief to bereavement, addressing a pressing need in end-of-life care for this vulnerable population.
Methods
Participants
This RCT was conducted from October 2024 to November 2025. The study population comprised family caregivers of patients with advanced cancer receiving hospice care. Family caregivers were defined as unpaid, legally/biologically related family members (e.g., spouse, adult child, or other family members such as grandparents/cousins) designated by the patient as primary caregiver and providing multidimensional support. In China, where nuclear family members are often unavailable due to the one-child policy, geographic mobility, or advanced spousal age, other family members were eligible, with patient nomination rather than self-report determining eligibility to ensure that enrolled caregivers occupied the principal role recognized by the patient. Inclusion criteria were age ≥18 years and the ability to provide written informed consent. Exclusion criteria included a known diagnosis of severe mental illness (e.g., psychotic disorder, severe cognitive impairment) or prior participation in a formal bereavement counselling or grief-focused psychological intervention in the preceding 12 months. This study was approved by the Ethics Committee of the affiliated Suzhou Hospital of Nanjing University Medical School (approval number: IRB202511005RI), and the trial was registered on the Thai Clinical Trials Registry (TCTR) (https://www.thaiclinicaltrials.org) (No. TCTR20260427005).
A total of 138 subjects were recruited, with 111 meeting the criteria. They were randomly assigned to 2 groups: the intervention group (n = 55) and the control group (n = 56). The flowchart is shown in Figure 1. Group assignment was conducted confidentially using sealed opaque envelopes by researchers not involved in the intervention implementation.
Flowchart of participants.
Figure 1 Long description
The flowchart begins with ′Assessed for eligibility (n = 138).′ It shows exclusions: ′Excluded (n = 27)′ with reasons including ′Not meeting inclusion criteria: Illiterate (n = 15),′ ′Had received relevant interventions (n = 2),′ and ′Mental instability (n = 10).′ Part icipants were ′Randomized (n = 111)′ into two groups: ′Allocated to structured grief counseling (n = 55)′ and ′Allocated to usual psychological materials (n = 56).′ Both groups start at ′Baseline (T0).′ For the intervention group, ′T1 pre-loss assessment′ includes ′Intervention group (n = 53)′ with ′Loss to follow-up before T1 (n = 2).′ Follow-up at ′T2 1 month after death′ shows ′Intervention group (n = 49)′ with ′Excluded: patient not deceased (n = 4).′ At ′T3 3 months after death,′ ′Intervention group (n = 48)′ with ′Excluded: Lost to follow-up (n = 1).′ Finally, ′Analysed (n = 48).′ For the control group, ′T1 pre-loss assessment′ includes ′Control group (n = 55)′ with ′Loss to follow-up before T1 (n = 1).′ Follow-up at ′T2 1 month after death′ shows ′Control group (n = 50)′ with ′Excluded: patient not deceased (n = 5).′ At ′T3 3 months after death,′ ′Control group (n = 47)′ with ′Excluded: Lost to follow-up (n = 3).′ Finally, ′Analysed (n = 47).′ The flowchart details the allocation and follow-up process for both groups in the study.
Control group
Received psychoeducational materials once at enrollment plus routine psychological support services delivered at the same frequency as the intervention group. The psychoeducational materials consisted of standardized booklets and handouts containing general information about grief reactions, self-care tips, and community resources, with no therapeutic counseling components. The routine psychological support services involved general listening and non-directive support without a structured grief counseling protocol.
Intervention group
Received 8 individual sessions of approximately 60 minutes each. Four sessions occurred during the anticipatory phase after patient enrollment in hospice care, and the remaining 4 were delivered within 3 months post-death, at weeks 2, 4, 8, and 12. Sessions took place in a quiet counselling room on the palliative ward or, when travel was difficult, via a secure video link. Therapists were clinicians or hospice care nurses who had completed a 2-day workshop on the treatment manual and attended monthly supervision. A written manual specified objectives, key techniques, example phrases, and homework tasks for each session to ensure consistent delivery across therapists. This manualized, individually structured grief counselling program was initiated during the patient’s terminal phase and continued for 3 months after death, with sessions scheduled at identical time points and frequency as the control group’s routine support. The program was designed for family caregivers of advanced cancer patients receiving specialist hospice care. Its theoretical foundation was cognitive behavioral theory, which conceptualizes complicated grief as sustained by insufficient loss integration, negative global beliefs, and maladaptive avoidance, and posits that these processes can be directly targeted through cognitive and behavioral techniques (Kustanti et al. Reference Kustanti, Fang and Linda Kang2021). The timing and content also followed clinical practice guidelines that recommend systematic psychosocial and bereavement support from hospice admission through early bereavement and drew on evidence from structured cognitive behavioral grief interventions for relatives of cancer patients (LaPlante et al. Reference LaPlante, Hardt and Maciejewski2024).
The program comprised 5 sequential modules (Rosner et al. Reference Rosner, Rau and Kersting2025): (1) Assessment and Case Formulation: In this initial module, therapists and family caregivers jointly developed an individual cognitive-behavioral formulation that linked caregiving stressors and loss-related triggers to automatic thoughts, emotions, physiological arousal, and coping responses. (2) Emotion Regulation and Safety: This module focused on psychoeducation about normal and complicated grief, training in paced breathing and brief muscle relaxation, and scheduling small valued activities that counteracted social withdrawal and inactivity (Groen et al. Reference Groen, Menninga and Cath2022). (3) Processing of the Loss: The loss processing module used detailed narratives, imaginal revisiting, and letter writing to help family caregivers approach avoided memories, test guilt-related appraisals, and develop more balanced views of responsibility. (4) Meaning Reconstruction and Identity: This module focused on clarifying the significance of the relationship, identifying values and roles that could be carried forward, and exploring realistic new roles in family and community life. (5) Future Planning and Continued Bonds: This final module encouraged setting concrete short-term goals and designing simple rituals or memory projects that linked remembrance with reengagement in life. Each session followed the same flow: opening and brief emotion check, review of homework and recent events, focused work on the weekly theme such as guilt, unfinished business or role transition, and a collaborative summary with specific between-session tasks. At the end of every session, the therapist carried out a brief risk assessment, asked directly about suicidal ideation or self-harm, and updated the safety plan when needed. The intervention schematic is shown in Figure 2.
Overview of the intervention and control conditions.
Measures
Assessments were conducted at 4 time points: at enrollment (T0, baseline), at the end of the intervention period (T1, defined as immediately following the patient’s death, which also marks the completion of the pre-loss phase of the intervention; note that post-loss sessions continued after this time point), 1 month post-death (T2), and 3 months post-death (T3). At each time point, family caregivers completed self-report questionnaires as described below.
Grief
The Anticipatory Grief Scale (AGS) was developed by Theut et al. at the National Children’s Medical Center in Washington, USA, in 1991, and revised into Chinese by Xin in 2016 (Theut et al. Reference Theut, Jordan and Ross1991). The scale comprises 7 dimensions: sadness, sense of loss, anger, irritability, guilt, anxiety, and ability to perform tasks, totaling 27 items. It employs a 5-point Likert scale, ranging from 1 (strongly disagree) to 5 (strongly agree). The total scores range from 0 to 135 points, with higher scores indicating more severe anticipated grief among family caregivers. The current study yielded a Cronbach’s alpha coefficient of 0.876.
Bereavement
Bereavement-specific guilt was measured with the Bereavement Guilt Scale (BGS), a multidimensional instrument developed and validated to assess several facets of guilt in the context of loss (e.g., self-blame and regret) (Li et al. Reference Li, Stroebe and Chan2015). We computed total and subscale scores following the developers’ scoring guidance (range 15–75). The higher the scale scores, the stronger the caregiver’s sense of guilt. The current study yielded a Cronbach’s alpha coefficient of 0.836.
Anxiety
Anxiety symptoms were assessed using Zung’s Self-Rating Anxiety Scale (SAS), a 20-item self-report measure rated on 4-point Likert responses (Zung Reference Zung1971). We summed item scores to yield a raw total (range 20–80) and transformed to the standardized index scores by multiplying by 1.25 (range 25–100), as per the original description. Higher scores indicate more severe anxiety symptoms. The current study yielded a Cronbach’s alpha coefficient of 0.849.
Depression
Depressive symptoms were measured with Zung’s Self-Rating Depression Scale (SDS), a 20-item instrument with 4-point Likert responses (Zung et al. Reference Zung, Richards and Short1965). Raw totals (20–80) were converted to standardized index scores (25–100) using the original algorithm. Higher scores indicate more severe depressive symptoms. The current study yielded a Cronbach’s alpha coefficient of 0.867.
Post-traumatic growth
Post-traumatic growth refers to positive psychological changes experienced as a result of struggling with highly challenging life circumstances, such as loss. Positive psychological change following loss was assessed with the Post-traumatic Growth Inventory (PTGI) (21 items; 5 domains: New Possibilities, Relating to Others, Personal Strength, Spiritual Change, Appreciation of Life). Items are rated 0–5; total scores range 0–105, with domain scores computed per the original specification (Tedeschi and Calhoun Reference Tedeschi and Calhoun1996). Higher scores indicate greater post-traumatic growth. The current study yielded a Cronbach’s alpha coefficient of 0.877.
Health-related quality of life
Health-related quality of life was conceptualized as the perceived physical and mental health status of caregivers, reflecting the impact of caregiving and bereavement on daily functioning and well-being. Health status of family caregivers was captured with the Short-Form 36 Health Survey (SF-36), covering 8 domains (physical functioning, role-physical, bodily pain, general health, vitality, social functioning, role-emotional, mental health) (Ware and Sherbourne Reference Ware and Sherbourne1992). Domain scores were transformed to a 0–100 metric. The higher the scale scores, the better the quality of life. The current study yielded a Cronbach’s alpha coefficient of 0.873.
Data collection
At baseline (T0), demographic and caregiving-history data (e.g., age, sex, relationship to patient, caregiving duration, hours per week) were collected. The randomization envelope was opened after baseline measurements. Follow-up assessments at T1 (end of intervention), T2 (1 month post-death), and T3 (3 months post-death) were scheduled and organized by the research team. Questionnaires were completed either in paper form or electronically (depending on participant preference). Research staff blinded to group allocation entered and managed the data; data entry was double-checked for accuracy. Missing data and drop‐outs were logged, and reasons for attrition recorded.
Statistical analyses
All statistical analyses were conducted using R version 4.3.2. Descriptive statistics were first computed to summarize participants’ demographic and baseline characteristics. To evaluate intervention effects across the 4 measurement points (T0–T3), linear mixed-effects models (LMMs) were applied for each continuous outcome (e.g., grief, guilt, anxiety, depression, post-traumatic growth, and quality of life).
In the model, group (intervention vs. control), time (T1–T3), and their interaction (group × time) were included as fixed effects, while a random intercept was specified for each participant to account for within-subject correlations due to repeated measurements. The baseline scores (T0) of each outcome were entered as a covariate to control for initial differences between groups. Additional covariates such as age and gender were included if imbalanced at baseline.
LMMs were chosen because they efficiently handle repeated-measures data with unequal time intervals and missing values under the missing-at-random assumption, thereby maximizing statistical power and minimizing bias (Bolker et al. Reference Bolker, Brooks and Clark2009). Estimated marginal means with 95% confidence intervals were derived using the emmeans package. Model assumptions were checked by residual diagnostics. Statistical significance was set at p < 0.05 (2-tailed).
Results
Descriptive analysis
The sociodemographic characteristics of both groups are presented in Table 1. No sociodemographic characteristics showed statistically significant differences between the 2 groups (P > 0.05).
Sociodemographic characteristics of participants
Table 1 Long description
Socio demographic characteristics are compared between an intervention group with 55 participants and a control group with 56 participants, with columns for a test statistic and P value. Mean age was 55.04 plus or minus 13.17 years in the intervention group and 54.07 plus or minus 12.43 years in the control group, with P 0.691. Sex counts were male 37 and female 18 in the intervention group, and male 33 and female 23 in the control group, with P 0.362. Educational level was middle school or below 29, high school 17, and high school or above 9 in the intervention group, versus 32, 13, and 11 in the control group, with P 0.647. Daily care duration in hours was less than 4 for 13, 4 to 12 for 29, and more than 12 for 13 in the intervention group, versus 11, 37, and 8 in the control group, with P 0.314. Total care time in months was less than 1 for 9, 1 to 3 for 19, and more than 3 for 27 in the intervention group, versus 14, 27, and 15 in the control group, with P 0.052. Duration of patient illness in years was less than 1 for 10, 1 to 3 for 16, and more than 3 for 29 in the intervention group, versus 17, 13, and 26 in the control group, with P 0.320. Relationship to the patients marked with an asterisk was son or daughter 34, spouse 17, parents 3, and other 1 in the intervention group, versus 31, 19, 2, and 4 in the control group, with P 0.586. Employment status was employed 23, unemployed marked with a number sign 0, and retired 32 in the intervention group, versus 24, 2, and 30 in the control group, with P 0.354. Annual household income was less than 50,000 RMB for 10, 50,000 to 100,000 RMB for 17, and more than 100,000 RMB for 28 in the intervention group, versus 6, 26, and 24 in the control group, with P 0.204.
Data are presented as mean ± standard deviation or n (%), as appropriate.
a Fisher’s exact test.
b Unemployed including homemakers, students, or other non-working status.
Scores on each scale at each time point
As shown in Figure 3, the corresponding scale scores for grief, guilt, anxiety, and depression in both groups generally declined over time, while the corresponding scale scores for post-traumatic growth and quality of life generally increased over time. The mean scores and standard deviations for each scale between the intervention group and the control group are presented in Figure 3.
Scores on each scale over time by group (mean ± SD). Subfigure (A) shows AGS scores, subfigure (B) shows BGS scores, subfigure (C) shows SAS scores, subfigure (D) shows SDS scores, subfigure (E) shows PTGI scores, and subfigure (F) shows SF-36 scores. AGS stands for the Anticipatory Grief Scale, BGS stands for the Bereavement Guilt Scale, SAS stands for the Self-Rating Anxiety Scale, SDS stands for the Self-Rating Depression Scale, PTGI stands for the Post-traumatic Growth Inventory, and SF-36 stands for the Short-Form 36 Health Survey.
Figure 3 Long description
Six line graphs display scores on each scale over time by group, mean plus or minus standard deviation. Each graph has a horizontal axis labeled Time point with three time points and a vertical axis representing scale scores. Two lines appear in each graph: one for the intervention group and one for the control group, distinguished by different line styles with markers and error bars representing standard deviation at each time point. Graph A: The vertical axis range appears to span approximately 0 to 100. The intervention group line starts at approximately 75 at time point 1, decreases to approximately 50 at time point 2 and continues declining to approximately 30 at time point 3. The control group line starts at approximately 60 at time point 1, decreases slightly to approximately 55 at time point 2 and drops to approximately 45 at time point 3. The intervention group shows a steeper decline overall. Graph B: The vertical axis range appears to span approximately 0 to 80. The intervention group line starts at approximately 70 at time point 1, decreases to approximately 35 at time point 2 and further declines to approximately 20 at time point 3. The control group line starts at approximately 65 at time point 1, decreases to approximately 50 at time point 2 and reaches approximately 40 at time point 3. Both lines decline with the intervention group declining more steeply. Graph C: The vertical axis range appears to span approximately 0 to 80. The intervention group line starts at approximately 30 at time point 1, rises to approximately 55 at time point 2 and increases further to approximately 70 at time point 3. The control group line starts at approximately 35 at time point 1, rises to approximately 45 at time point 2 and reaches approximately 55 at time point 3. Both lines increase over time. Graph D: The vertical axis range appears to span approximately 0 to 80. The intervention group line starts at approximately 65 at time point 1, decreases to approximately 40 at time point 2 and declines to approximately 20 at time point 3. The control group line starts at approximately 60 at time point 1, decreases to approximately 50 at time point 2 and reaches approximately 40 at time point 3. Both lines decline steadily. Graph E: The vertical axis range appears to span approximately 0 to 80. The intervention group line starts at approximately 70 at time point 1, decreases to approximately 35 at time point 2 and reaches approximately 15 at time point 3. The control group line starts at approximately 65 at time point 1, decreases to approximately 50 at time point 2 and reaches approximately 35 at time point 3. Both lines decline with the intervention group showing a sharper drop. Graph F: The vertical axis range appears to span approximately 0 to 80. The intervention group line starts at approximately 20 at time point 1, rises to approximately 50 at time point 2 and reaches approximately 70 at time point 3. The control group line starts at approximately 25 at time point 1, rises to approximately 35 at time point 2 and reaches approximately 45 at time point 3. Both lines increase over time with the intervention group rising more steeply. A legend present in each graph identifies the intervention group and control group lines.
Results of LMM analysis
The results of the LMM analysis are shown in Table 2. The results revealed significant differences between groups in AGS scores (P < 0.001), with the intervention group exhibiting lower AGS scores than the control group (estimates = −9.31 and −7.79, respectively). Thus, over time, the intervention group maintained lower AGS scores compared to the control group, a finding further supported by the lower mean scores for the indicator in the intervention group shown in Figure 3. Additionally, the intervention group scored higher than the control group on the PTGI and SF-36 (estimates = 6.02 and 5.97, respectively; P < 0.001). Thus, over time, the intervention group exhibited higher post-traumatic growth and better health-related quality of life than the control group. Figure 3 also shows that the intervention group had higher mean PTGI and SF-36 scores compared to the control group. On the other hand, differences in BGS scores, SAS scores, and SDS scores between groups were not significant over time (all P > 0.05).
Results of linear mixed-effects model analysis (interaction terms)
Table 2 Long description
Interaction term results are listed for six parameters with columns for Estimate, SE, t, P, and 95%CI with lower and upper bounds. BGS scores show an estimate of 0.94, SE 0.92, t 1.019, P 0.310, with a 95%CI from minus 0.88 to 2.76. SAS scores show an estimate of minus 0.65, SE 1.03, t minus 0.634, P 0.527, with a 95%CI from minus 2.69 to 1.38. SDS scores show an estimate of 0.34, SE 1.03, t 0.336, P 0.737, with a 95%CI from minus 1.69 to 2.38. PTGI scores show an estimate of 6.02, SE 0.41, t 14.680, P less than 0.001, with a 95%CI from 5.21 to 6.83. SF 36 scores show an estimate of 5.97, SE 0.41, t 14.477, P less than 0.001, with a 95%CI from 5.16 to 6.78.
AGS stands for the Anticipatory Grief Scale; BGS stands for the Bereavement Guilt Scale; SAS stands for the Self-Rating Anxiety Scale; SDS stands for the Self-Rating Depression Scale; PTGI stands for the Post-traumatic Growth Inventory; SF-36 stands for the Short-Form 36 Health Survey.
Discussion
This RCT tested whether a structured, CBT-based grief counseling program delivered from the terminal phase through 3 months post-loss improves psychological outcomes in family caregivers of patients with advanced cancer. In intention-to-treat analyses that adjusted for baseline values, the intervention produced statistically significant reductions in grief severity over time and yielded higher post-traumatic growth and better health-related quality of life. Between-group differences for guilt, anxiety, and depression were not significant. Taken together, these findings indicate that a time-limited, structured protocol aligned to the anticipatory and early bereavement trajectory can meaningfully affect grief-specific outcomes while supporting broader adaptation.
The pronounced reduction in grief scores within the intervention group is a cornerstone finding of this study. The significant group-by-time interaction (estimate = −9.31, P < 0.001) demonstrates that the intervention actively facilitated the grieving process over time, leading to a lower level of complicated grief symptomatology compared to the control group. Our grief findings are consistent with prior randomized trials of structured, manualized, family-focused therapies initiated during hospice care and extended into bereavement (Kustanti et al. Reference Kustanti, Fang and Linda Kang2021). Family-Focused Grief Therapy, for example, reduced complicated-grief severity versus usual care and lowered the incidence of prolonged grief disorder in at-risk families, supporting the value of anticipatory grief work that continues post-loss. Mechanistically, the structured grief counseling program likely reduced maladaptive avoidance and dysregulated appraisals that maintain grief distress; both exposure to loss-related cues and cognitive restructuring are central change processes identified in randomized trials of Complicated Grief Therapy and in cognitive-behavioral models of complicated grief (Kissane et al. Reference Kissane, McKenzie and Bloch2006). Positioning sessions across both the pre-loss and early post-loss windows may also have supported adaptive oscillation between loss-oriented and restoration-oriented coping, a core prediction of the Dual Process Model that aligns with family-systems approaches (Palm et al. Reference Palm, Doering and Kubiak2023).
The positive association between the present intervention and post-traumatic growth is consistent with evidence that meaning-oriented or growth-focused programs can facilitate growth following loss. A randomized trial of a manualized meaning-making program for bereaved adults, as well as early internet-based cognitive-behavioral grief treatments, demonstrated gains in post-traumatic growth (Roepke et al. Reference Roepke, Tsukayama and Forgeard2018). More specifically, Meaning-Centered Grief Psychotherapy (MCGP), developed by Lichtenthal and colleagues to target meaning and purpose in bereavement, has been shown to enhance spiritual well-being and meaning, with effects corresponding closely to the PTGI domains of new possibilities and personal strength (Lichtenthal and Breitbart Reference Lichtenthal and Breitbart2015; Applebaum et al. Reference Applebaum, Roberts and Lynch2022). Our structured CBT-based program shares conceptual foundations with MCGP, particularly in its emphasis on meaning reconstruction and identity work. It differs, however, in its phased delivery spanning both the pre-loss and post-loss periods, as well as its broader integration of cognitive-behavioral techniques such as exposure and behavioral activation. These meaning-centered strategies are theorized to foster constructive reappraisal, clarified values, and a sense of continuity with the deceased-pathways that align with PTGI domains and are supported by trials of meaning-centered psychotherapy in advanced cancer populations (Tedeschi and Calhoun Reference Tedeschi and Calhoun1996). Consistent with this framework, the observed improvement in health-related quality of life parallels meta-analytic and trial evidence indicating that structured, skills-based caregiver interventions can yield small to moderate gains in quality of life for families facing cancer (Chow et al. Reference Chow, Mathews and Cheng2023).
Conversely, the lack of significant intervention effects on guilt, anxiety, and depression warrants careful consideration. First, the nonsignificant guilt finding suggests that while the intervention addressed yearning and emotional pain of grief, it may not have specifically targeted the cognitive ruminations and self-blame characterizing guilt, which often roots in specific death circumstances and may require more focused cognitive techniques. Second, bereavement-related guilt includes distinct cognitive facets (e.g., responsibility guilt) that may resist change unless explicitly targeted with cognitive restructuring or exposure. Third, the 3-month follow-up may have been insufficient to capture guilt changes, which can persist or intensify in early bereavement before declining. Fourth, regarding anxiety and depression, while the intervention included emotion regulation and behavioral activation, these symptoms may require more intensive or sustained intervention, such as additional sessions focused specifically on mood management. Meanwhile, grief-focused interventions tend to exert their largest and most consistent effects on grief-specific outcomes, with smaller and more heterogeneous effects on anxiety and depression across trials and follow-up intervals (Hao et al. Reference Hao, Qiu and Liang2024).
This trial has several strengths, including a randomized group design with allocation concealment, initiation of counseling in the pre-loss phase with continuation into early bereavement, and the use of multiple validated psychological scales. However, limitations include a single-center setting and a moderate sample size that may reduce power for secondary outcomes. Second, follow-up ended at 3 months after the death, which may be insufficient to detect slower changes in certain negative emotional states. Third, prior bereavement history was not collected and may confound responses to loss. Future research should be multi-site with larger and more diverse samples and longer follow-up.
Implications
These findings support the integration of structured grief counseling programs into intervention systems for bereaved family caregivers. The structured grief counseling, bridging the pre-loss and early bereavement periods, should be systematically implemented to alleviate negative emotions such as complex grief among family caregivers. Furthermore, the intervention’s promotion of post-traumatic growth and health-related quality of life highlights its role in fostering adaptive psychological adjustment. Services should screen family caregivers at enrollment and at transition points for grief severity, post-traumatic growth, quality of life, and care needs, and use these data to triage intensity. A stepped care approach is indicated, adding guilt-focused cognitive strategies and mood modules when standard counseling does not reduce residual distress. Clinical protocols should thus prioritize this continuous, manualized support to enhance both psychological well-being for family caregivers navigating the transition through anticipatory grief and bereavement.
Conclusion
In this RCT, a manualized CBT-based grief counseling program delivered from the terminal phase through 3 months after the death produced meaningful benefits for family caregivers of patients with advanced cancer. Compared with usual care, participants receiving the intervention showed a sustained reduction in grief severity and gains in post-traumatic growth and health-related quality of life. No between-group differences emerged for guilt, anxiety, or depression, indicating that grief-focused protocols exert their strongest effects on grief-specific outcomes while mood and guilt may require additional targeted strategies.
Competing interests
The authors declare no competing interests.
Open access
