The experience of cancer caregiving may have adverse effects on the psychosocial health of family caregivers; these effects might persist after the family member's death and could lead to various health problems.1,2 Findings from studies investigating these effects are ongoing and continue to demonstrate a variety of serious consequences, for example, high rates of psychological distress,3 postloss depressive symptoms and anxiety symptoms,4 and prolonged grief disorder.5 It has also been demonstrated that between 40% and 76% of cancer patients' caregivers experience sleep disorders.6 The occurrence of sleep disturbances in bereavement has been confirmed in various studies.7 The health of bereaved cancer caregivers is, therefore, of continuing concern.
Delivering interventions targeted at reducing the risk of these adverse consequences on these caregivers might be beneficial. The importance of continuity of supportive care from palliative services by offering bereavement aftercare has been emphasized.8 Bereaved family caregivers have reported that they are not adequately supported or prepared emotionally for their loss.9,10 In their systematic review, Nielsen et al9 found that a caregiver's lack of preparedness was associated with poorer bereavement outcomes after the death of a close relative and might also lead to a higher risk of developing complicated grief symptoms.11 A recent study10 demonstrated that caregivers who reported lower levels of preparedness had higher levels of anxiety and depression. The findings also identified that caregivers who were more anxious and depressed during caregiving experienced more adverse consequences post loss. The authors point out that their results indicate that having adequate support, both emotional and practical, during the caregiving phase might have a positive impact on bereavement outcomes. However, there are many aspects of bereavement that emerge after a death of a family member because of cancer, and the most effective modes of intervening remain underresearched and need to be explained and measured more accurately.
A multicomponent Family Strengths-Oriented Therapeutic Conversation (FAM-SOTC) intervention has recently been evaluated and was developed with the focus on providing psychosocial support to family caregivers of a close relative in the final phases of cancer receiving specialized palliative care.12 Family caregivers participating in this intervention trial reported, among other positive outcomes, significant reduction in stress symptoms after receiving 2 sessions of the intervention. The authors emphasized the importance of acknowledging the emotional responses of family caregivers and giving them space to express their worries and concerns through this type of therapeutic conversations.12 An extended postloss version of the FAM-SOTC has also been developed based on the theoretical frameworks of the Calgary Assessment and Intervention Models.13 In a recent review of family caregiving intervention trials in oncology, it was reported that future studies should translate tested supportive-based models for caregivers into clinical practice.14 This study is a trial intervention where participants were offered a third session of the FAM-SOTC intervention, specifically the extended version targeting the postloss experience. The family caregivers had received 2 sessions of the intervention before the death of their close relative because of cancer.
In addition, to reach a greater understanding of the bereaved caregivers' grief reactions after a loss of a family member because of cancer, we used the Range of Response to Loss Model (RRLM) developed by Machin15 as a theoretical framework. Machin15 identified 3 main reaction patterns to loss: those who are "Overwhelmed" (eg, experiencing emotional distress of grief that may interfere with everyday functioning), those who are "Controlled" (eg, in denial of or with the need to suppress distress), and those who are "Resilient" (eg, able to cope with and accept strong feelings of grief as part of the loss experience). These RRLM concepts were used to create a self-assessment tool, the Adult Attitude to Grief (AAG) scale.15 This instrument is used in the current study. Studies about specific preloss and continuing postloss supportive nursing interventions and their impact on bereavement reactions are scarce. This study addresses this knowledge gap by describing the effect of providing ongoing support provided by a palliative nurse to family cancer caregivers in the period after the death of a close relative.
The study objectives were to investigate the impact of the intervention on bereaved family caregivers' psychological distress (depression, anxiety, and stress symptoms) and grief reactions (eg, overwhelmed, controlled, and resilient) when 2 sessions of the FAM-SOTC intervention have been provided during the advanced/final stage of cancer pre loss and 1 session is offered post loss and to compare these outcomes to those following care as usual.
On the basis of the theoretical frameworks of the Calgary Assessment and Intervention Models,13 which guided the development of the FAM-SOTC intervention and review of the literature, we hypothesized that bereaved family cancer caregivers who had received 3 sessions of the FAM-SOTC intervention (eg, 2 sessions pre loss and 1 session post loss) would report significantly lower levels of psychological distress (depression, anxiety, and stress) compared with those receiving usual care, measured 3, 5, and 6 months after loss. The decision for these time points is based on findings supporting a relatively high prevalence of depressive and anxiety symptoms among this population 2 to 6 months after the patient's death because of cancer.4 The secondary aim based on the theoretical frameworks of the RRLM15 was to answer the following research question: are there differences in reactions to loss among bereaved caregivers in the intervention group who have received the FAM-SOTC intervention pre and post loss and those in the control group who received usual care pre and post loss, measured at 3, 5, and 6 months after loss?
METHODS
The study design was quasi-experimental, with a posttest-only comparison of the intervention and control groups that were recruited from a single-center hospital-based specialized palliative home care unit. The posttest-only design was used in this case because control group participants were not evaluated before the loss, because they were not given the intervention and were only recruited after the death of their family member.
Sample and Setting
The inclusion criteria were that the primary family caregivers were older than 18 years and bereaved because of cancer and that 3 months had passed since the patient had died. An additional eligibility criterion was that the family caregivers had received services from the palliative home care unit in the advanced and final stage of the illness.
Fifty-one bereaved family caregivers were assessed for eligibility, and all met criteria and were invited to participate. Twenty-six of these had initially participated in an earlier evaluation of the intervention; these formed the intervention group. The other 25 participants were not exposed to the intervention and formed the control group. None of the participants in the intervention or control group were lost to follow-up.
Demographic Data
In the full sample, most were female (67%; Table 1). There was a significantly higher percentage of women in the control group compared with the intervention group (84% vs 54%, respectively, with P = .02). Participants were most frequently 61 years and older (63% of the total sample). Most had an elementary or high school education (73%), and the bereaved family caregivers were mostly spouses (80%). There was no difference between the intervention and control groups on age, education, relationship of caregiver, or the number of family members living in the home.
Control Group
Bereaved family caregivers in the control group received usual care, for example, 1 home visit before answering the questionnaires at the first time point. Usual care post loss comprised an unstructured conversation, and the palliative home care nurse provided information about community services. No clinical practice guidelines are used in the unstructured conversation; for example, the way in which usual care is provided varies according to the individual nurse.
Intervention Group
The FAM-SOTC intervention focuses on supporting the cognitive, affective, and behavioral domains of the family member's illness experience.12 The theoretical core of the intervention is based on the Calgary models from the family systems nursing models. Findings from previous family-based studies support the use of therapeutic interventions and narrative approaches to bereavement care in which the bereaved are given the opportunity to talk about their preloss and postloss experiences.16 Offering bereaved individuals to express verbal narratives after the death of a significant other can be perceived as helpful in making sense of their loss.17
The adapted FAM-SOTC postloss intervention comprises the following 5 core components, drawn from the Calgary models: (1) eliciting narratives about the preloss and postloss experience; (2) asking therapeutic/interventive questions, emphasizing the most pressing concerns, and using therapeutic listening; (3) validating/acknowledging emotional responses; (4) assessing the need for specific information and recommendations regarding bereavement; and (5) the use of commendation/focusing on the strengths of the bereaved caregiver (Table 2).
The bereaved family caregivers in the intervention group (n = 26) received 1 face-to-face session of the FAM-SOTC postloss intervention at their home. The intervention was provided by an advanced practice palliative care nurse and lasted for approximately 60 to 90 minutes. These same participants had also received 2 FAM-SOTC intervention sessions in an earlier trial12 before the loss of their family member. The first author collected all the data from the participants in both the intervention and control groups at their homes. There was an opportunity for a follow-up assessment at the end of the study (at 6 months post loss) as well as the option to refer those in need to appropriate health care services. A detailed description of the intervention implementation session in clinical practice, the rationale for each of the 5 components and which core domain of the family caregivers' illness experience the intervention is targeting, is presented in Table 2. A detailed manual was also presented when reporting the preloss trial, and it is recommended to provide the first 2 sessions of the intervention when the patient and his/her family are entering the palliative care services. Further participation in the therapeutic conversation intervention can then be offered at any other time point with the focus on supporting specific care needs and reducing negative effects of the vulnerable situation.
Instruments
All participants in both groups answered the same questionnaires at 3, 5, and 6 months post loss. Baseline bereaved family caregiver demographic data collected included sex, age, education, relationship to the deceased, and the number of family members in the home (Table 2).
The Depression Anxiety Stress Scale (DASS) was used to assess the psychological distress of participants.18 The DASS is a validated and reliable 42-item self-reported instrument with 3 subscales (depression, anxiety, and stress), each comprising 14 items. A 4-point Likert reaction scale is used for respondents to indicate the extent of each emotional state they have experienced for the past week. In this study, the Cronbach [alpha] values for depression, anxiety, and stress were 0.89, 0.84, and 0.91.
The AAG scale15 was used in this study to measure the validity of grief reactions, ranging from overwhelmed feelings and controlled functioning and including resilient capacity in balancing these factors.15 The AAG is a validated and reliable self-report questionnaire, which contains 3 statements for each of the Overwhelmed, Controlled, and Resilient categories from the RRLM, which was developed by the same author to explain individuals' grief reactions and coping responses to a loss.15 The AAG scale has the potential to show individual changes in reactions to loss over time and, according to the author, can be used as a method for enhancing the understanding of bereaved individuals' expressions and experiences of grief.15 The AAG consists of 9 items, and higher scores on the 5-point Likert-type response scale indicate greater severity of grief reactions. The Cronbach [alpha] values for Overwhelmed, Controlled, and Resilience in this study were 0.67, 0.54, and 0.61. It has been demonstrated in the literature that a lower coefficient can be expected (eg, 0.5) when subscales have only 3 items (fewer than 10) in each.19
Outcome Measures
The primary outcomes of the current study were the 3 subscales of the DASS measure of psychological distress. Depression, anxiety, and stress were measured at 3 postloss time points: 3, 5, and 6 months after the death of the family member.
Secondary outcomes were the core domains of grief reactions, measured using the AAG scale. Consistent with the study design, Overwhelmed, Controlled, and Resilient outcomes were also measured at 3, 5, and 6 months post loss.
Ethical Approval
Informed consent was obtained from bereaved family caregivers in the control group by the palliative care nurses and by the nurse researcher for caregivers in the intervention group. Approval for the study was obtained from the Scientific Ethical Board at Landspitali the National University Hospital of Iceland (no. 50/2013). The study was reported to the National Data Protection Authority (no. S6569/2013) and registered at http://Clinicaltrails.gov (ID: NCTO3850132).
Data Analysis
Study variables were summarized using means and standard deviations or frequency distributions. Group comparisons were conducted with 2-sample t tests, [chi]2 tests of association, or Fisher exact tests. Treatment group comparisons over the postintervention period were made using mixed models for repeated measures. These models included the main effects of group (treatment vs control) and time (3, 5, and 6 months) as well as their interaction; sex was included in the models as a covariate because there was a difference in sex distribution between the groups. Post hoc analysis for significant effects was performed using Fisher least significant difference procedure for pairwise comparisons. When using a post hoc test, the results of the experimental data of the study can be analyzed further, and a Fisher least significant difference tool is used in this study to identify which pairs of means are statistically different. Data analysis was completed with SAS version 9.4 (SAS Institute, Inc, Cary, North Carolina), with an [alpha] of .05 for inferential tests. This study is an evaluation of a subset of participants who were included in a treatment-only study to evaluate the intervention over time.
RESULTS
Mixed Modeling for Primary Outcomes
The model with the DASS Depression subscale was significant overall, but the only model effect that was close to the .05 level of significance was time (P = .052; Table 3). This result shows that there is a difference in both groups at any of the 3 time points. Post hoc pairwise evaluation of this effect suggested that the 5-month average depression level in the intervention and control participants as a group exceeded that of the 6-month average. This result indicates that depression levels at 5 months among participants in the intervention and control groups combined are higher at 5 months than at the 6-month time point. This relationship is consistent with the means shown in Figure 1.
Both group and time were significant in the model based on the Anxiety subscale of DASS (P = .048 and P = .0045, respectively; Table 3). This result shows that there is a difference between groups at any of the 3 time points. The post hoc analysis demonstrates that the intervention participants had lower scores, on average, compared with the control participants (P = .048; Table 3). In addition, the 6-month mean was less than both 3- and 5-month means for this outcome (P = .0048 and P = .0034, respectively; Table 3). These relationships are reflected in Figure 2.
For the outcome of DASS Stress, only the time main effect was significant (P = .034). The post hoc analysis for this outcome also indicates that the 6-month mean was less than both 3- and 5-month means (P = .037 and P = .025, respectively; Table 3). The Group x Time interaction and the sex covariate were not significant in any of the models for these primary outcomes.
These results suggest that the hypothesis of improved psychological distress outcomes after the loss of a family member because of cancer, among those participating in the intervention, was partially supported: intervention group participants had significantly lower anxiety scores compared with those in the control group throughout the posttest assessment period (3, 5, and 6 months), for example, after having received the intervention.
Mixed Modeling for Secondary Outcomes
The repeated-measures mixed modeling for the secondary outcomes of AAG subscales exhibited fewer significant effects than the DASS models. Although the models for Overwhelmed and Resilient reaction patterns were significant overall, none of the model effects were significant for either model. For the Controlled outcome, the main effect of group was significant indicating that individuals with this dominant reaction pattern may be in denial or suppressing their distress. The intervention participants had lower scores for this subscale, on average, compared with those in the control group (P = .043; Table 3).
DISCUSSION
This study evaluates the impact on psychological distress symptoms (depression, anxiety, and stress) and grief reactions after loss when providing an extended FAM-SOTC intervention to family caregivers after the death of their close relative because of advanced cancer. We found that the intervention was associated with lower levels of anxiety symptoms among treatment participants relative to controls in the same postloss period. This result suggests, in line with the study hypothesis, that the FAM-SOTC intervention did have a sustained positive impact on this indicator of psychological distress symptoms among family caregivers. In particular, the findings revealed a significant decrease in anxiety symptoms in the intervention group compared with the control group across all 3 time points. Anxiety and stress symptoms also decreased over time in the 2 groups combined. The intervention did not improve depressive symptoms, although there was a trend toward decreasing depression over time for the 2 treatment groups combined, particularly between 5 and 6 months.
The findings of the secondary outcomes indicate that caregivers in the control group seemed to have a dominant "Controlled" reaction to loss. Individuals experiencing this grief reaction pattern are more likely to be in denial of or have a need to suppress distress.15 This dominance of a controlled pattern15 might indicate that these reactions to loss could be having a negative impact on everyday functioning and emotional well-being.
This intervention study has several key insights that may be relevant for practice; it seems that the relationship established with family members through a therapeutic conversation before the death of a close relative may have critical implications for future development of supportive interventions targeted to bereaved cancer caregivers.20 Establishing an effective relationship with patients and family caregivers has been reported as one of the essential components of palliative care.21 Zech et al22 have suggested that aspects of the therapeutic relationship might have the potential to affect outcomes of bereavement interventions. They have pointed out that the quality of the therapeutic relationship might also be of importance regarding these outcomes. The positive effects of the FAM-SOTC intervention might be partly the result of the empathic,23 nonjudgmental attitude essential for developing a therapeutic relationship when this specific family-based approach is used to empower family members in coping and should be studied further (Table 2). The importance of providing caregiver interventions in clinical settings or in the home, with a focus on family care, has recently been emphasized.24 Offering the bereaved caregiver continuing participation in a therapeutic conversation after the loss of a close relative because of cancer might mitigate intense suffering25 and the negative consequences after bereavement.16,17 The intervention efforts can be tailored to the unique care needs of the bereaved and presents the opportunity to assess mental health. As frontline palliative home care providers, nurses and those in hospice could foster more effective bereavement coping skills through the mechanism of a therapeutic conversation intervention presented in Table 2 and identify those in need for specific services and resources. Although the DASS tool is used for the purpose of evaluating the impact of the intervention, other instruments should be used when applying the intervention in clinical practice. The National Comprehensive Cancer Network Distress Thermometer measuring both anxiety and depressive symptoms is recommended when providing the intervention.26
The results of the current study underscore the importance of assessing and supporting family caregivers' specific needs of care post loss and providing an extended brief supportive intervention continuing into bereavement.
Strengths and Limitations
The core strength of this study is its use of theoretical guidance to develop the postloss intervention; this conceptual basis provided insight into complex aspects of the bereavement period. Another strength is the complete retention of participants over the 6-month postloss period, with no participant withdrawal in either study group. High dropout rates have been reported as a potential reason for the ineffectiveness of supportive care interventions for family cancer caregivers.27 The primary limitation of this study is the nonrandomized design, given that the intervention participants had already been recruited for the parent study. This design may have been part of the reason that sex distributions in the 2 groups were different. Widows are commonly overrepresented in bereavement studies generally, which is the case in the control group for this study. The sex balance was more even in the intervention group, perhaps because all the widowers from the larger preloss trial who were invited to participate in this secondary study agreed to do so. Most of the bereaved family caregivers in this study were female spouses and older than 60 years. Intrapersonal risk factors, for example, sex, age, and nature of the relationship, may influence outcomes after bereavement. Previous studies have demonstrated that both young and old ages are linked to high levels of grief symptoms.28 Findings from a recent study of 748 older adults exploring sex differences in mental health conditions after widowhood indicated that widows were better able to cope with loss and are more resilient than widowers.29 However, findings from a population-based study of bereaved partners and nonpartners of terminally ill patients found no sex difference in levels of grief symptoms measured at 6 months post loss.30 This study is single-sited, and the number of eligible participants was finite; the relatively limited sample size may have blunted our ability to detect the full impact of the FAM-SOTC postloss intervention on caregiver bereavement outcomes. Future studies will benefit from a randomized design, with a sample size planned to attain a prespecified level of power.
This study informs nurses about the beneficial effect of providing a brief nursing intervention in reducing psychological distress among bereaved family cancer caregivers. The study may promote nurses in providing this supportive nursing intervention to family cancer caregivers receiving palliative home care.
CONCLUSION
This study demonstrated that a FAM-SOTC nursing intervention resulted in reducing psychological distress symptoms among bereaved family caregivers and that the degree of anxiety and controlled reactions post loss may be particularly affected by exposure to this brief intervention. Bereaved family cancer caregivers might therefore benefit from being offered a theory-driven FAM-SOTC intervention.
Acknowledgment
The authors thank all the bereaved family caregivers who participated in this study and the palliative care nurses at the specialized palliative home care unit for their valuable contribution by assisting with recruitment.
References