Variations in grief, anxiety, depression, and health among family caregivers before and after the death of a close person in the context of palliative home care

Abstract

This article investigates longitudinal variations in grief, self-rated health, and symptoms of anxiety and depression among family caregivers in palliative care. Data were taken from a randomized psycho-educational intervention trial and were collected at four time-points; at baseline, upon completion, 2 months later, and 6 months after the patient’s death. In total, 117 family caregivers completed all questionnaires. The participants’ grief was stable across the measurements, while anxiety, depression, and health varied significantly (p < 0.05). No significant differences were found between the intervention or control group. In conclusion, grief emerged as a constant phenomenon, distinct from symptoms of anxiety and depression.

Family caregivers are often essential in home-based palliative care and may face many care-related tasks and have to take on great overall responsibility (Morris, King, Turner, & Payne, 2015). Caring for and supporting a close person with incurable illness has been described as a double experience with both emotional challenges and rewards (Anderson & White, 2018b). On the one side, caregiving can be described as meaningful and can contribute to personal development. However, caregivers also describe feelings of being insufficiently prepared and supported (Anderson & White, 2018a), resulting in negative consequences, both emotionally and socially (Lund, Ross, Petersen, & Groenvold, 2014). Family caregivers may develop sleeping problems across the patient’s disease trajectory (Lee, Hsieh, Lin, & Lin, 2018) and also have a greater risk of developing mental health problems compared to general populations (Shaffer et al., 2017).

Family caregivers not only need to prepare for specific aspects of caregiving, such as practical assistance, medical tasks, and providing emotional and existential support, but they must also prepare for the expected death. Preparing to care for a close person with severe illness may in itself contribute to a feeling of becoming aware of death, which can be frightening, but also helpful, providing a sense of control (Henriksson & Andershed, 2007; Holm, Henriksson, Carlander, Wengstrom, & Ohlen, 2014; Janze & Henriksson, 2014). Further, awareness of death can be beneficial for long-term well-being (Hauksdottir, Steineck, Furst, & Valdimarsdottir, 2010). Still, the loss of a close person to death is considered one of the most challenging losses in life (Zisook et al., 2014). Although a natural and unavoidable part of life, grief could spark various forms of poor health (Jensen, O’Connor, Spindler, & Moskowitz, 2018; Zisook et al., 2014). Persistent depressive symptoms have been observed in those who are bereaved, which usually diminishes in frequency and intensity over time (Kuo, Sun, & Tang, 2017). Anxiety, due to the loss of a close person, and being confronted by ones’ own mortality, is also common in bereavement (Shear & Skritskaya, 2012).

However, bereavement may not be the starting point to the grieving process, as the grief could begin long before a close person’s death; when the patient is diagnosed with an incurable illness, during the progression of the disease, and at the end of life. In a recent review, pre-death grief (or anticipatory grief) was associated with negative consequences in bereavement (Nielsen, Neergaard, Jensen, Bro, & Guldin, 2016). The authors concluded that interventions should focus on supporting family caregivers during pre-death grief to develop coping strategies to deal with losses encountered during caregiving and to increase family caregivers’ feelings of preparedness, both in their caregiving and for the subsequent bereavement.

There has been a substantial increase in the number of interventions aiming to support family caregivers in various ways (Candy, Jones, Drake, Leurent, & King, 2011; Chi, Demiris, Lewis, Walker, & Langer, 2016; Harding, List, Epiphaniou, & Jones, 2012). Supportive interventions during palliative care could have a positive effect on outcomes, not only during caregiving, but also post-bereavement (Aoun, Ewing, Grande, Toye, & Bear, 2017). Several studies on psycho-educational interventions have been successful in improving the preparedness for caregiving in family caregivers, both in short- and long-term (Henriksson, Arestedt, Benzein, Ternestedt, & Andershed, 2013; Hudson et al., 2013, 2015). Earlier research has also suggested that family caregivers who feel prepared and knowledgeable also experience fewer problems both pre- and post-bereavement (Hebert, Prigerson, Schulz, & Arnold, 2006; Nielsen et al., 2016), however, this needs to be explored further.

To gain more knowledge on how to support family caregivers in palliative care, there is a need for further studies on the variations of grief and the possible negative consequences related to grief, such as anxiety, depression, and poor health in family caregivers, both pre- and post-bereavement. It must also be investigated whether providing support during caregiving could influence these variables.

Aim

The aim of this study was to investigate longitudinal variations in grief, symptoms of anxiety and depression, and self-rated health among family caregivers in palliative home care.

Methods

Design

This explorative study used longitudinal data which were derived from a randomized control trial (RCT) in palliative home care, developed in accordance with the CONSORT statement (Boelen & Smid, 2017). The trial was registered at www.clinicaltrials.gov/ (ID: NCT02482415). The trial had two arms: a psycho-educational intervention in three sessions, and a control group with standard care support. The aim of the intervention was to increase preparedness for caregiving and a previous study showed that it was effective in this (Holm, Årestedt, Carlander, Fürst, Wengström, Öhlen & Alvariza, 2016). Although the focus of this paper was not to evaluate the intervention, it was decided to include possible differences between the intervention and control group in the result, as the intervention could still have influenced the family caregivers who participated. Data were collected at 4 time-points; baseline, upon completion of the intervention, 2 months after intervention, and 6 months after the patient’s death. The intervention was delivered by health care professionals in palliative care, but the authors of this study had the overall responsibility for the project, including intervention development, randomization, and data collection. Approval was granted from the regional ethical review board in Stockholm, Sweden (2012/377-31, 2012/2191-32, 2013/934-32).

Sample and procedure

Altogether, 10 specialized palliative home care facilities in a metropolitan area in Sweden were invited to participate in the trial. All facilities agreed to participate. Each facility provided advanced palliative care to a range of 70–200 patients in their own home. The facilities mainly enrolled patients diagnosed with cancer, but also those with coronary, pulmonary, or neurological diseases.

Patients and family caregivers were invited to participate by health care professionals at the palliative care facilities. Initially, the patient was approached and had to provide consent, and, if he or she did, the family caregiver was approached and given study information and a request to participate. The inclusion criteria were: being a family caregiver to a person in specialized palliative home care, over the age of 18, and able to read and understand the Swedish language.

Data were collected through questionnaires, which consisted of self-rated instruments and background questions, including age, sex, social status, and relationship to the patient. The baseline questionnaire was given to the family caregivers when they agreed to participate. When the baseline questionnaire was completed and sent to the authors of this study, the family caregiver was block randomized either to the intervention or the control group (standard support). Randomization was based on blocks of four, using a computer randomized number generator, and stratified for the 10 home care settings. Family caregivers randomized to the intervention were posted an invitation to the three sessions.

The second, third, and fourth follow-up questionnaires were sent by post to family caregivers by the authors of this study. The fourth follow-up was sent depending on the time of the patient’s death and, hence, the timing could vary greatly. The questionnaires were returned in a prepaid envelope. The intervention took place between 2013 and 2014. The final follow-up assessment was concluded in January 2016.

The intervention

The intervention has been thoroughly described elsewhere (Holm, Carlander, Fürst, Wengström, Årestedt, Öhlen & Henriksson, 2015) and hence, only a brief description will follow. The intervention had a psycho-educational design and was developed in steps based on the theoretical framework of Andershed and Ternestedt (Andershed & Ternestedt, 2001) relating to family caregivers’ principal needs for knowledge and support. The framework was the basis for an intervention manual, which was used by health care professionals (physician, nurse, and social worker/priest) as a framework for three intervention sessions spread over 3 weeks.

Measurements

The questionnaires consisted of background variables and validated instruments. The Anticipatory Grief Scale (AGS) was used to measure the variations in pre-death grief at baseline and follow-ups 1 and 2. However, if the patient had died before the next follow-up, family caregivers were instructed not to complete the AGS, because it presupposes that the patient is still alive. The Texas Revised Inventory of Grief (TRIG) was used to measure post-death grief at the third follow up, 6 months after the patient’s death. The Hospital Anxiety and Depression Scale (HADS) and the Health Index (HI) were used to measure variations in health and symptoms of anxiety and depression in all four measurements.

The AGS was originally created to measure the intensity of pre-death grief in family caregivers of patients with dementia (Theut, Jordan, Ross, & Deutsch, 1991). The AGS is partly based on TRIG and several items in the two scales are equivalent. For this study, the AGS-13, a newly validated, revised 13-item version of the AGS was used, which had satisfactory psychometric properties (Holm, Alvariza, Fürst, Öhlen & Årestedt, 2019). The AGS-13 is validated for the context of palliative home care and measures pre-death grief over two scales; Behavioral reactions, and Emotional reactions, which have 8 and 5 items, respectively. All items are answered on a 5-point Likert-type scale, where higher values indicate higher levels of pre-death grief. The total score of each scale ranges between 8 and 40, and 5 and 25, respectively. Cronbach’s alpha for both scales was 0.83 in the present study, indicating satisfactory internal consistency.

The TRIG measures the intensity of grief after the death of a close person (Faschingbauer, 1981). It consists of 21 items divided into two scales. TRIG I (Past Feelings) has 8 items and involves thinking back to the time just after the loss. The TRIG II (Present Behaviors) consists of 13 items and measures the current situation in bereavement. The TRIG is measured on a 5-point Likert-type scale where lower values indicate higher levels of grief. The two subscales are summed into a total score, ranging from 8 to 40 for TRIG I, and from 13 to 65 for TRIG II. The instrument has satisfactory psychometric properties for a palliative care context (Holm, Alvariza, Fürst, Öhlen & Årestedt, 2018).

The Hospital Anxiety and Depression Scale (HADS) consists of 14 items divided into two subscales measuring symptoms of anxiety and depression, respectively. Each item has four response alternatives ranging between 0 and 3, where higher values indicate more anxiety and depression. Each subscale has a total score between 0 and 21. In this study, HADS has been used as a continuous scale, however, cutoff scores have been suggested at 0–7 = no case, 8–11 = possible case, and 11–21 = cases (Bjelland, Dahl, Haug, & Neckelmann, 2002). Cronbach’s alpha was 0.87 and 0.86 for the two subscales in the present study. The psychometric properties of the Swedish version have been found to be satisfactory (Djukanovic, Carlsson, & Arestedt, 2017; Lisspers, Nygren, & Soderman, 1997).

The Health Index consists of 11 items measured on a 1–4 scale, where higher values indicate better health. The total score ranges between 11 and 44. Cronbach’s alpha for the present study was 0.85. The scale has satisfactory psychometric properties (Rohani, Khanjari, Abedi, Oskouie, & Langius-Eklof, 2010).

Analysis

Data were validated before analysis by an independent person, and missing values were replaced if they did not exceed 20% through person mean imputation (Allison, 2001). In total, 51 missing values were replaced. Statistical analyses were performed with SPSS statistics version 24 (IBM corp., Armonk, NY, USA). A p-value of <0.05 was considered statistically significant.

Descriptive statistics were used to describe the sample of family caregivers. Characteristics included participants’ age, gender, social status and relation to the patient. It also included their baseline ratings of the outcome variables. To test for baseline differences between the intervention and control group at baseline, independent t-tests or Pearson chi-square tests were used.

A mixed design analysis of variance (ANOVA) was performed to investigate variations in the outcomes over time (pre-death grief, anxiety, depression, and health), group differences (intervention group and control group) and the interaction effect of time and group. The four time-points (baseline, follow-ups 1, 2, and 3) were used as the within-subjects factor, and group allocation (random or control) was used as the between-subjects factor. Post-hoc analyses were conducted through pairwise comparisons between the scores at the various time points, adjusted with Bonferroni correction for multiple comparisons. Independent sample t-tests were performed to investigate the influence of the intervention on post-death grief (measured only at follow-up 3).

Results

Characteristics

In total, 117 family caregivers completed the measurements at baseline and the two follow-ups (Table 1). A majority were family caregivers to a patient diagnosed with cancer, although heart and pulmonary, and neurological diseases were also represented. On average, 4 years had passed since the patient received his or her diagnosis. Significant differences were observed between the intervention and the control group at baseline with regards to age, where the intervention group was significantly older (p = 0.045), and social status, where the intervention group was more often married or in a partnership (p = 0.037). No significant differences were found between the intervention group or the control group with regards to their pre-death grief, anxiety, depression or health at baseline. The length of time between the intervention and the patient’s death ranged between <1 month to >2 years.

Table 1. Demographic characteristics and differences in characteristics between control and intervention group.

Baseline characteristics	All participants (n = 117)	Control group (n = 59)	Intervention group (n = 58)	p-value
Age, mean (SD)	62 (13.1)	59 (14.7)	64 (10.8)	0.045^a
Gender n (%)				0.277^b
Women	75 (65.1)	35 (59.3)	40 (69.0)
Men	42 (35.9)	24 (40.7)	18 (31.0)
Social status n (%)				0.037^b
Married/partnership	89 (76.1)	40 (67.8)	49 (84.5)
Unmarried	28 (23.9)	19 (32.2)	9 (15.5)
Relation to patient n (%)				0.080^b
Spouse	59 (50.4)	23 (39.0)	36 (62.1)
Adult child	39 (33.3)	25 (42.4)	14 (24.1)
Other	19 (16.3)	11 (18.6)	8 (13.8)
Education level n (%)
Academic degree	54 (46.2)	29 (49.2)	25 (43.1)	0.572^b
Non-academic degree	63 (53.8)	30 (50.8)	33 (56.9)
Variables, mean (SD)
AGS13-beh	19.2 (6.3)	19.2 (6.3)	19.3 (6.3)	0.763^a
AGS13-emo	14.3 (5.2)	14.4 (5.4)	14.2 (5.1)	0.483^a
HADS-anxiety	7.4 (4.2)	7.1 (4.4)	7.7 (4.0)	0.580^a
HADS-depression	5.2 (3.3)	5.2 (3.5)	5.2 (3.1)	0.852^a
HI	33.7 (4.7)	33.5 (4.7)	33.9 (4.7)	0.495^a

^a Independent sample t-test;

^b Chi-square test.

Longitudinal variations in pre-death grief, anxiety, depression, and health

No longitudinal variations (p = 0.386 and p = 0.383) or group differences between the intervention or control group (p = 0.757 and p = 0.332) were shown for pre-death grief, measured by AGS-13 behavioral and emotional reactions, respectively. Also, no interaction effect of time and group were detected for behavioral (p = 0.974) or emotional (p = 0.982) reactions (Table 2).

Table 2. Variations over time, between groups and interaction effect of time and group based on a mixed design ANOVA.

Outcomes	Baseline, mean (SD)	Follow-up 1, mean (SD)	Follow-up 2, mean (SD)	Follow-up 3, mean (SD)	Source^a	df	F-value	p-value	Post-hoc test^c
AGS13-beh (n = 117)	19.3 (6.3)	19.5 (6.4)	19.8 (6.8)		Time	2.0	0.96	0.386	–
Control group	19.2 (6.3)	19.6 (5.6)	19.7 (6.4)		Allocation	2.0	0.28	0.757
Intervention group	19.3 (6.3)	19.3 (6.8)	19.9 (7.1))		Allocation × Time	1.0	0.00	0.974
AGS13-emo (n = 117)	14.3 (5.2)	14.3 (5.3)	14.0 (5.4)		Time	2.0	0.96	0.383	–
Control group	14.4 (5.4)	14.5 (5.5)	14.5 (5.5)		Allocation	2.0	1.10	0.332
Intervention group	14.2 (5.1)	14.1 (5.2)	14.2 (5.1)		Allocation × Time	1.0	0.00	0.982
HADS-anxiety (n = 115)	7.4 (4.2)	7.8 (4.3)	7.5 (4.5)	6.2 (3.9)	Time	3.0	11.1	<0.001	- - - D E F
Control group	7.1 (4.4)	7.5 (4.3)	7.5 (4.5)	6.1 (4.1)	Allocation	3.0	0.71	0.545
Intervention group	7.7 (4.0)	8.3 (4.3)	7.5 (4.5)	6.4 (3.8)	Allocation × Time	1.0	0.37	0.544
HADS-depression (n = 115)	5.2 (3.3)	6.0 (4.2)	6.2 (4.0)	5.6 (4.1)	Time^b	2.7	3.80	0.014^b	- B - - - -
Control group	5.2 (3.5)	5.7 (4.2)	6.3 (4.2)	5.7 (4.4)	Allocation	2.7	0.62	0.584
Intervention group	5.2 (3.1)	6.3 (4.1)	6.1 (3.9)	5.5 (3.8)	Allocation × Time	1.0	0.01	0.941
HI (n = 115)	33.7 (4.7)	32.9 (4.9)	32.3 (5.0)	32.5 (5.2)	Time^b	2.8	5.90	0.001^b	AB - D - -
Control group	33.5 (4.7)	32.8 (4.6)	32.0 (4.9)	32.4 (5.3)	Allocation	2.8	0.32	0.798
Intervention group	33.9 (4.7)	32.9 (5.2)	32.8 (5.1)	32.7 (4.9)	Allocation × Time	1.0	0.22	0.643

^a Time - main effect within subjects over time. Allocation - main effect between intervention and control, Allocation × Time - interaction effect for time and allocation;

^b Greenhouse-Geisser corrected p-values due to violations of sphericity;

^c Pairwise comparisons. Significant differences (p < 0.05) are presented as; A = Baseline-Follow-up 1; B = Baseline-Follow-up 2; C = Follow-up 1-Follow-up 2; D = Baseline-Follow-up 3; E = Follow-up 1-Follow-up 3; F = Follow-up 2-Follow-up 3. AGS = Anticipatory Grief Scale; HADS = Hospital Anxiety and Depression Scale; HI = Health Index.

Both symptoms of anxiety and of depression, measured by HADS, demonstrated significant longitudinal variations (p < 0.001 and p = 0.014). There was a significant difference between the three pre-death measurements of anxiety and the post-death measurement, where the levels of anxiety significantly decreased. For depression, there was a significant variation between the baseline measurement and the second follow-up, where the levels of depression increased. No significant group differences between the intervention or control groups (p = 0.545 and p = 0.584, respectively), or interaction effects for anxiety or depression (p = 0.544 and p = 0.941, respectively) were demonstrated (Table 2).

Self-rated health also demonstrated significant longitudinal variations (p < 0.001). The ratings of health decreased significantly between baseline and the first, second, and third follow-up. No significant group differences between intervention or control groups (p = 0.798), or interaction effects (p = 0.643) were demonstrated (Table 2).

Post-death grief

Seventy-eight family caregivers completed the final measurement on post-death grief, 33 in the control group, and 45 in the intervention (Table 3). There were no significant differences between the intervention group and the control group, either for Past behaviors (TRIG I) (p = 0.979) or Present feelings (TRIG II) (p = 0.901).

Table 3. Differences between the control group and the intervention group on post-death grief measured by the Texas Revised Inventory of Grief (TRIG).

Outcome	Total sample (n = 78) Mean (SD)	Control group (n = 45) Mean (SD)	Intervention group (n = 33) Mean (SD)	p-value
TRIG-I	30.0 (1.0)	30.1 (1.3)	29.8 (1.4)	0.902
TRIG-II	38.0 (1.4)	38.9 (1.9)	36.8 (2.2)	0.462

Discussion

From the results found in this study, no longitudinal variations in family caregiver’s ratings of grief before death were found, while their ratings of anxiety, depression, and health varied significantly over the four measurement points before and after the death of the patient. No differences were found between the two groups (intervention and control), either before or after the death of the patient.

Family caregivers’ ratings of pre-death grief were stable across the three time-points at which these were measured, indicating no variations over time, both with regards to behavioral reactions and emotional reactions. Both the pre-death and post-death ratings indicate that family caregivers’ feelings of grief in this study were moderately intense. However, this is only in comparison to the total score, as there are no established clinically meaningful values for either of these scales. It should be remembered that the sample in this study consisted of family caregivers to a patient who received palliative care, which has been described as a task that demands full attention, 24 hrs every day. The pressure of this situation could delay, interrupt, or in other ways influence the process of grief (Holtslander et al., 2017), which could explain the stable and moderate ratings of grief. It is also possible that the care setting for this study (specialized palliative home care), which involves frequent visits from health care professionals, might have had an ameliorating impact on family caregivers, as earlier results have demonstrated that support from professionals close to home could promote family caregivers’ feelings of resilience, which is a protective factor during caregiving (Roen et al., 2018). Further, the fact that the time the diagnosis was made was sometimes months or even years before the study began, could also mean that family caregivers had time to become aware of the patient’s condition and future death, which is associated with less intense feelings of grief and less problems in bereavement (Hauksdottir et al., 2010; Kim, Carver, Spiegel, Mitchell, & Cannady, 2017; Valdimarsdottir, Helgason, Furst, Adolfsson, & Steineck, 2004).

Family caregivers’ post-death grief was measured 6 months after the patient’s death. The moderate levels of post-death grief could be an indicator that the participants had been given time to adjust to the loss and were in an integrated rather than an acute state of grief (Zisook et al., 2010). Although grief is not something that dissolves with time (Arizmendi & O’Connor, 2015), this integrated form of grief is described as being less intense than the initial, acute grieving phase. Integrated grief includes the ability to understand and accept the death, return to life, function adequately, and the ability to look ahead towards future life in a positive way (Zisook et al., 2014).

Unlike their grief, family caregivers’ symptoms of anxiety and depression varied significantly over the four measurement points. These different patterns add to a body of research suggesting that anxiety and depression are phenomena that are distinct from grief (Boelen, 2013; Dillen, Fontaine, & Verhofstadt-Deneve, 2009). The level of anxiety decreased significantly after the patient’s death, which has also been observed in earlier studies (Petruzzi et al., 2015; Waldrop, 2007). The levels of depression also decreased after the patient’s death, although not significantly. However, depression increased gradually between the three measurements pre-death, and there was a significant increase between the first and the third measurement. This is in line with earlier results, showing that family caregivers often exhibit higher levels of anxiety and depression during caregiving as the patient’s condition deteriorates (Williams & McCorkle, 2011), and a substantial decline after the patient’s death (Gotze et al., 2018; Schulz et al., 2003). Qualitative results have demonstrated that family caregivers may feel powerless and invisible during caregiving (Linderholm & Friedrichsen, 2010), which could possibly contribute to increased feelings of anxiety and depression, and which might subside after the patient’s death. Although symptoms of anxiety and depression varied over time, the levels were moderate, particularly for depression, over the four measurements.

For health, significant variations were observed between the baseline and the three follow-ups. Health levels decreased over the measurements pre-death, with a slight improvement after the patient’s death. Family caregivers often have unmet needs, such as those required for communication and practical support, which could explain their decreasing health over the caregiving trajectory. Many caregivers are quite old and frail themselves, and this can make it difficult to participate in practical caregiving, such as washing or feeding the patient (Morris et al., 2015).

The fact that family caregivers both from the intervention and control group in this study improved their ratings in anxiety, depression, and health after the patient’s death could indicate that their situation during palliative caregiving was stressful and challenging compared to their situation post-death, which has been well-documented in earlier studies (Antony, George, & Jose, 2018). It could also indicate that they were quite adequately prepared for the loss, as preparedness has been highlighted as a factor that could indicate fewer problems post-death (Hebert, Dang, & Schulz, 2006). A clinical implication of these results could be that family caregivers need support during caregiving to manage symptoms of anxiety and depression and health problems. There are scientifically developed ways to help caregivers directly identify and prioritize their own needs and this could be used by health care professionals as a means to focus on family caregivers’ over the patient’s illness trajectory (Ewing, Grande, & Home, 2013)

Limitations

It should be noted that patients in this study were assigned to specialized palliative care facilities that promote a holistic approach focused on symptom relief, quality of life, and the provision of support, both for the patient and the family caregivers, regardless of whether they participated in the intervention or not (Garcia-Perez, Linertova, Martin-Olivera, Serrano-Aguilar, & Benitez-Rosario, 2009). For this reason, it is possible that the levels of and variations in grief, anxiety, depression, and health would have been different in another context.

This study used two different instruments to measure grief. Although AGS is partly based on TRIG, it includes several questions that are not equivalent to the ones in TRIG. Using the same instrument, both pre- and post-death would have made it possible to evaluate the variations in grief across all four measurement points. Further, TRIG was only measured on one occasion, 6 months after the patient’s death. To explore variations in post-death grief, future studies should include a measurement taken closer to the patient’s death, where the intensity of grief could have been higher, and also a measurement later into bereavement.

Although the intervention described in this study was not designed to reduce grief, it should be highlighted that the intervention could have influenced family caregivers’ experiences of grief. For this reason, it was considered important to include group differences (intervention and control) in the analysis of variance. Rather than using a one-way repeated measures ANOVA, we used a mixed model, where both the time-points and group allocation were included. This was considered to be a way to be certain that the intervention had no influence on the grief of the participants included in the study.

Conclusion

In conclusion, no longitudinal variations between family caregivers’ ratings of grief over time were found which might indicate that it is possible to manage and integrate grief into the continued life. The variations in anxiety, depression, and health showed different patterns than grief and varied significantly over time. Thus, the study affirms that grief is to be regarded as a phenomenon that is distinct from symptoms of anxiety and depression. This study further adds to the knowledge about the stressful situation that family caregivers experience when caring for a close person in the context of palliative care. This can be concluded as their levels of anxiety and depression decreased after the death of the patient, and their levels of self-rated health improved.

Disclosure statement

The authors declare no conflicting interests.

Additional information

Funding

This work was supported by the Swedish Cancer Society (Cancerfonden).