Abstract
Objectives
Persistent Complex Bereavement Disorder (PCBD) is an independent clinical condition characterized by a severe grief reaction lasting more than twelve months after the loss of a loved one. This study aimed to assess the rate of PCBD among caregivers of end-of-life patients. The secondary aim was to explore the role of several risk factors for PCBD, including psychosocial characteristics of both patients and caregivers, patients’ dignity-related distress, attachment style, and perceived social support as well as symptoms of depression, anxiety, and perceived stress of caregivers.
Methods
This was a prospective, multicenter and national study involving twelve Italian palliative care units. The sample consisted of 139 cancer patient–caregiver dyads. For each participant a set of rating scales was administered.
Results
In line with the literature, only a minority of caregivers experienced PCBD (6.47%). Younger caregiver age and higher pre-loss depression levels significantly predicted PCBD six months after the patient’s death.
Introduction
Loss is a profoundly painful experience, forcing an individual to adapt to a life forever altered. 1 Grief occurs when a deep emotional bond is severed, leaving those who remain in profound distress. 2 After a loss, an initial acute pain response occurs. Usually, this phase lasts weeks or months before eventually ending in the so-called integrated grief, with suffering gradually diminishing within 6-12 months. 3 Although most bereaved people follow this trajectory, research articles report an incidence or prevalence rate – which varies greatly – of individuals who develop Complicated Grief (CG), a condition in which grief persists and intensifies. 4
According to the Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition (DSM-5), 5 people who have such a severe reaction to the loss of a loved one lasting longer than 12 months should be diagnosed with Persistent Complex Bereavement Disorder (PCBD). Individuals with this disorder may experience a profound sense of suffering and are unable to accept the death resulting in significant emotional and functional impairment. In addition, they may experience identity confusion as though they have lost a part of themselves. 6 Their symptoms include separation distress, i.e., longing and searching for the person who has died, preoccupation with the deceased, and symptoms of traumatic distress, i.e., feelings of disbelief, anger, shock, and detachment from others. 7
Common causes contributing to PCBD can be divided into pre-loss risk factors (i.e., age of the caregiver/patient, prior loss, insecure attachment, previous history of depression, and social support)7-12 and loss-related factors (i.e., closeness and intensity of the relationship with the deceased, anxiety and depression disorders, nature of the death itself, resources available after death).11-13
Although existing research has primarily focused on grief mediators in the general bereaved population, 14 in recent years several studies have specifically examined the experiences of caregivers in palliative care. Among these, a review by López Pérez et al, (2022) determined the factors associated with complicated grief in caregivers of deceased persons in palliative care; the variables that have been investigated as main risk factors are: age of the deceased, dedication to care, relationship with the deceased, education level, financial aspects, anticipatory grief intensity, depression symptoms or mental disorders of caregivers, affective dependency and maladaptive attachment styles. 11 Another study by Coelho et al (2016) stood out for having explicitly investigated the grief mediators in a Portuguese sample of caregivers of patients who had died in palliative care, who were more than 6 months bereaved. 15 They found that relational factors, such as insecure and dependent relationships, unresolved family crisis and perceived deterioration and disfigurement of the patient were associated with PCBD. Regarding the studies conducted in Italy, Nanni et al assessed family members of terminally ill patients through the Inventory of Complicated Grief (ICG) to identify subjects at risk for PCBD but without exploring risk factors. 16 In contrast, Lai et al explicitly evaluated the determinants of PCBD in a sample of caregivers of end-of-life patients, finding that caregiver burden was the best predictor of PCBD. 17
Several external factors can contribute to the development of PCBD in caregivers.
For instance, the pain experienced by terminally ill cancer patients can significantly shape the grief process of their caregivers. 3 Indeed, end-of-life patients can be affected by dignity-related distress due to the inability to perform everyday life tasks (washing or dressing), psychological symptoms (anxiety, depression, fatigue), and physical symptoms (nausea, pain, breathing difficulties). 18 This distress can alter patients’ perceived sense of dignity, affecting not only their quality of life, but also that of their caregivers, with significant repercussions on the process of separation.19,20 Therefore, the previously unexplored hypothesis underlying the present study was that beyond the factors related to caregivers, patient dignity-related distress may play a role in predicting caregiverss’ PCBD.
The primary aim of the present study was to measure the rate of PCBD in a sample of caregivers of end-of-life patients, 12 months after the death of their loved one. Corresponding to our hypothesis, the secondary aim was to evaluate the role of several risk factors for PCBD, including psychosocial characteristics of patients and caregivers (i.e., patient age, patient education, caregiver’s age, and type of patient-caregiver relationship), patient’s dignity-related distress, attachment style, and perceived social support as well as symptoms of depression, anxiety, and perceived stress of caregivers.
Materials and methods
This was a prospective, multicenter and national study involving twelve Italian palliative care units. These palliative care units were identified through the regional coordinators of the Italian Palliative Care Society (SICP) and through the network of psychologists and psychotherapists working in palliative care.
The study involved three observation time points: T0, the admission of the patient in the palliative care service (hospital, hospice, nursing home, home care); T1, six months after the patient’s death; and T2, twelve months after the patient’s death. The mean time interval between T0 and T1 was 195 days (SD = 15.8).
Study Population
Between September 2021 and April 2023, a consecutive sample of patient-caregiver dyads was recruited, where each dyad consisted of a patient with terminal cancer admitted to a palliative care unit and their primary caregiver.
Inclusion criteria for patients were the following: ≥18 years old; having a diagnosis of cancer; being able to provide informed consent; and meeting the criteria to access palliative care according to the Italian National Law on Palliative Care and Pain Treatment No. 38/2010, 21 i.e., having an advanced cancer without any appropriate curative treatments, with less than four months of estimated life expectancy, and scoring 50 or lower on the Karnofsky Performance Status (KPS). 22
Regarding caregivers, inclusion criteria were: ≥18 years old; being able to provide informed consent; having a close relationship with the end-of-life patient.
Exclusion criteria for patients and caregivers included: having a diagnosis of a severe psychiatric disorder and/or cognitive impairment with an inability to provide informed consent or to complete the protocol study.
All participants received psychological support as part of their routine care. During the first consultation, the psychologist explained the research. All participants agreed to take part. Participating patients completed the self-report scale and sociodemographic information. Afterwards, for each patient, his/her caregiver was contacted by phone by a psychologist because, due to COVID-19 restrictions, external visitors were not allowed access to the palliative care service. In this way, caregivers were asked to provide sociodemographic information, and they were administered rating scales. Caregivers were also contacted and assessed by phone at six and twelve months after the patient’s death.
All participants were informed about the aims of the research and provided written informed consent.
The study was approved by the Local Ethical Committee (protocol No. 0097316, procedure No. 403/2021, 23/09/2021). The authors state that the research was conducted in accordance with the latest version of the principles of the Declaration of Helsinki.
Assessment Tools
Demographic and sample characteristics including patient and caregiver age, gender, education in years, type of caregiver-patient relationship (partner, adult child, other relative) were collected via self-report at study baseline. The patient’s dignity and the caregiver variables (i.e., levels of depression, anxiety, stress, social support, attachment style, and grief) were assessed using the following tools:
The Patient Dignity Inventory 23 (PDI), Italian version,24,25 is a 25-item questionnaire that measures various sources of distress related to patients’ dignity. It is divided into 5 subscales:
Psychological Distress, Social Support, Physical Symptoms and Dependency, Existential Distress, and Loss of Purpose and Meaning. The items are on a 5-point Likert scale ranging from 1 (not a problem) to 5 (an overwhelming problem).
The Depression Anxiety Stress Scales-21 26 (DASS-21), Italian version, 27 is a 21-item scale that measures the level of depression, anxiety, and stress in the past week. The items are on a 4-point Likert scale, from 0 (never) to 3 (always).
The Medical Outcomes Study-Social Support Survey 28 (MOS-SSS), Italian version, 29 is a 20-item scale that assesses perceived social support. It is divided into 4 subscales: Emotional Support, Tangible Support, Affective Support, and Positive Social Interaction. Each item (except the first one) is rated on a 5-point Likert scale from 1 (never) to 5 (every time).
The Experience in Close Relationships Scale 12 30 (ECR-12), Italian version, 31 is a 12-item questionnaire that measures attachment style to the romantic partner in adults. It is divided into 2 subscales: Avoidant Attachment and Anxious attachment. The items are on a 7-point Likert scale, from 1 (strongly disagree) to 7 (strongly agree).
The Prolonged Grief Disorder Questionnaire 32 (PG-12), Italian version, 33 is a 12-item questionnaire that describes feelings, thoughts and actions experienced by the caregiver of a terminally ill patient. Items are divided into three parts: the first one assesses how often, in the past month, the caregiver experienced emotions related to their loved one’s illness (from 1 = never, to 5 = several times a day); the second part investigates how much a patient’s illness affected the caregiver’s life, (from 1 = never, to 5 = unbearable); the third part consists of a single question, dichotomously answered (yes/no), which explores whether there is a significant reduction in social, work, or other important areas of the caregiver’s life.
The Prolonged Grief Disorder-13 34 (PG-13), Italian version 35 (LP-13), is a 13-item questionnaire divided into three parts. Items investigate feelings, thoughts, and actions that have persisted for six months since the loved one’s loss and that are associated with a significant functional individual impairment.
Procedures
At T0, patients completed the PDI. At the same time point, caregivers completed four measures: DASS-21, MOS-SSS, ECR-12, PG-12. At T1 and T2, caregivers completed the DASS-21 and the PG-13.
Statistical Analysis
All computations were performed using RStudio for macOS (Version 2024.04.0 + 748, RStudio Inc, Boston, MA, USA).
The multivariate imputation by chained equations with Breiman’s random forest algorithm (number of imputations = 100, number of iterations = 10, and number of trees = 10) was adopted to address the issue of missing values due to dropout at follow-ups.
Comparisons between groups (PCBD vs non-PCBD) in terms of socio-demographic and clinical characteristics at baseline were carried out using Fisher’s exact test for the categorical variables, and the exact analogue of Wilcoxon–Mann–Whitney test for continuous variables.
The predictive value of the relevant variables for PCBD diagnosis at T1 and at T2 was then estimated by binomial generalized linear models (GLMs) with a stepwise selection of variables based on the Akaike Information Criterion (AIC).
For both the endpoints, the baseline model included only the predictors of PCBD already reported in the previous studies; it provided the significant predictors to be evaluated in the final model along with variables which turned out to be differently distributed in two groups at baseline in the univariate analyses. The center variable was also included in all the models to address a possible inter-center variability.
A P-value of 0.05 was used to designate statistical significance, but P-values resulting from multiple comparisons were adjusted by Šidák’s correction to control for the family-wise error rate.
Results
Sociodemographic Data of the Sample
Descriptive Characteristics of the Patients (N = 139)
Descriptive Characteristics of the Caregivers (N = 139)
The dropout rate was 5.8% (8 out of 139 participants) at T1 and 9.4% (13 out of 139 participants) at T2 (i.e., 5 caregivers at T2 and 8 at both T1 and T2). Caregivers who dropped out at both T1 and T2 could not be reached using the contact information provided, whereas those who dropped out at T2 specifically requested to withdraw from the study.
Patient-caregiver relationships included partners (36.7%, n = 51), adult children (33.8%, n = 47) and other relatives (29.5%, n = 41). Overall, 6.5% (n = 9) of caregivers met the diagnostic criteria for PCBD.
Assessment of the Predictors of PCBD (T1)
Six-Month Follow-Up Assessment (T1)
Selected: including literature predictors.
Final: Significant predictors of Selected model + Significant predictors resulting from univariate analysis.
Assessment of the Predictors of PCBD (T2)
Twelve-Month Follow-Up Assessment (T2)
Selected: including literature predictors.
Final: Significant predictors of Selected model + Significant predictors resulting from univariate analysis.
Discussion
The present study aimed to investigate the rate of PCBD developed by caregivers of patients who died of cancer, twelve months after the death of their loved one, and to explore also the possible risk factors.
In line with previous research suggesting that most caregivers adaptively process loss, with only a minority developing PCBD,7,35-37 this multi-center study found a prevalence of 6.47% . It is possible that the prospective nature of the study, combined with the additional attention given to caregivers’ feelings and sense of support during follow-up, may have influenced their bereavement experience, potentially contributing to the relatively low prevalence observed.
However, the results of the present study did not confirm the previously reported influence of psychosocial factors, perceived social support, and symptoms of depression and anxiety on the likelihood of developing PCBD one year after the loss, 14 as none of the caregiver characteristics examined in this study were identified as predictors of PCBD. This discrepancy could be related to differences in the study design, statistical methodology, and follow-up duration. To the best of our knowledge, this is the first multicenter study that has prospectively investigated predictors of PCBD across three observation time points, using a binomial GLM. Although the multicenter design may have introduced a recruitment bias, this was largely addressed by including inter-center variability in all statistical models which had negligible impact on the results. Furthermore, a strength of the present study was the use of multivariable statistical models to evaluate the predictors of PCBD, which enabled a more rigorous assessment of independent associations.15,16
Boelen et al used a prospective design similar to ours. Nevertheless, they examined the linkage of socio-demographic and loss-related variables, assessed within the first year of the bereavement with probable PCBD, assessed one year later, in the second year of bereavement. 38
Previous studies have shown that the most severe symptoms tend to occur in the first six months after death, when grief is most intense.12,16 Consistently, in the present study caregiver’s young age emerged as a predictor of PCBD at six months post-loss. In the literature, other studies have explored this aspect, with mixed results. In particular, López Pérez et al 11 showed that caregiver’s age is not a predictor of PCBD. Boelen et al 38 reported that probable PCBD was more prevalent among older caregivers. Despite differences in mean age between samples, the estimated age ranges for those at risk appeared comparable.
Moreover, pre-loss depression emerged as another predictor of PCBD six months post-loss. Caregivers who reported depressive symptoms at T0 presented grief symptoms consistent with PCBD at T1, as assessed using the PG-13 scale; here, PCBD refers to the presence of symptom clusters rather than a formal diagnosis. This result is consistent with findings by Lai et al, 17 who confirmed the central role of depression in the development of PCBD, potentially due to emotional and cognitive vulnerabilities that hinder adaptive grief processing. Conversely, in a recent French multicenter study caregivers’ depression symptoms during the end of life of their loved one were independent factors of PCBD. 12 At this point, it is necessary to note that psychological symptoms related to grief and depression can overlap. However, in PCBD, distress is primarily focused on feelings of loss and separation from a loved one, rather than reflecting a generalized low mood as in depression. 39 This distinction supports the decision to recognize PCBD as a separate diagnostic entity. 40
In contrast to prior evidence, attachment style to the romantic partner and perceived social support, were not found to be significant for PCBD. These findings are counterintuitive both in relation to the initial hypotheses and the existing studies in the literature, which suggest that insecure attachments in childhood and/or in adulthood with the romantic partner, are predictors of PCBD.12,41 One possible explanation is that grief responses are highly individual, shaped by resilience, life experiences, and other psychological resources. Thus, even caregivers with insecure attachment or low perceived social support may possess alternative coping mechanisms that help protect them from the development of prolonged grief.
Finally, patient dignity-related distress was not found to be a predictor of PCBD in caregivers. This may reflect the inherently subjective nature of dignity: although patients may experience distress over perceived loss of dignity, caregivers may not interpret or internalize this in a way that significantly impacts their grieving process.
Identifying risk factors during the end-of-life phase is crucial for timely, targeted interventions to prevent PCBD. Two promising approaches are Meaning-Centered Grief Therapy (MCGT), a therapeutic model for addressing challenges in finding meaning frequently faced by bereaved people, and Family Focused Grief Therapy, an intervention aimed at preventing complicated bereavement by enhancing family functioning.42,43 Overall, the findings of the present study emphasize the complexity of the grieving process and highlight the importance of early assessment.
Limitations and Insight for Future Research
The research has several limitations
Recruiting participants was difficult because of the COVID-19 pandemic: many hospices were converted into COVID wards, resulting in a decrease in bed availability and, consequently, a delay in patients’ access to palliative care units. As a result, the sample size was reduced. The low prevalence of PCBD in the sample may also have limited the identification of significant predictors. Further research with a larger sample may yield different findings.
Moreover, the study was performed in twelve Italian palliative care units; however, the non-homogeneous distribution of the sample limits the generalizability of the results. In addition, the findings may not be generalizable to other countries due to cultural differences.
Another potential limitation of the study is the absence of standardized or validated measures assessing the quality of the relationship between patients and caregivers. Future research including such measures could offer deeper insight into the relational factors that increase the risk of developing PCBD.
Furthermore, the advanced average age of both patients and caregivers is a limiting factor. It would therefore be useful to conduct a study focused on younger patients and caregivers, as grief may have different impacts depending on the age at which it occurs.
Conclusions
The present research showed that only a minority of caregivers developed PCBD.
These findings underline that younger caregiver age and pre-loss depression are predictors of PCBD symptoms six months after the patient’s death. The present study, supported by a longitudinal multicentric prospective design, highlights the importance of adopting a more careful approach to the assessment of risk factors for PCBD.
Footnotes
Acknowledgments
The authors thank all the caregivers and patients for participating in the study.
Ethical Considerations
The study was approved by the Local Ethical Committee (protocol No. 0097316, procedure No. 403/2021, 23/09/2021).
Consent to Participate
All participants were informed about the aims of the research and provided written informed consent.
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.
Declaration of Conflicting Interest
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Data Availability Statement
The data that support the findings of this study are available from the corresponding author, upon reasonable request.