Perceived burden and need for support among caregivers of cancer patients

Abstract Background The aims of this study were to examine (1) the perceived burden among caregivers and identify those in risk of high burden and (2) the need for support among caregivers and identify associated factors. Materials and methods Cancer patients who participated in an educational program at the Montebello Center (MBC) in Norway between May 2021 and February 2022 were asked to invite a caregiver to answer a questionnaire. The caregiving burden was assessed with the Caregivers Reaction Assessment (CRA) that consists of 24 questions scored from 1 (strongly disagree) to 5 (strongly agree), covering five domains. A mean sum score was calculated for each domain. Higher subscale scores indicate higher levels of burden, except for caregiver esteem where a high score indicates a low burden. Need for support was assessed with 13 questions. Results Of 464 invitations, 185 caregivers responded (response 40%), median age was 58.0 years and 58% were male. Caregiver burden mean scores were: 2.6 (SD 1.03) for Impact on schedule, 2.1 (SD 0.79) for Lack of family support, 2.1 (SD 0.76) for Impact on health, 2.0 (SD 0.86) for Impact on finances, and 4.2 (SD 0.47) for Caregiver esteem. Female caregivers, younger, higher education, having comorbidities, caring for patients having recurrence of cancer, and shorter time since diagnosis were associated with higher burden measured on individual subscales of the CRA. Most reported needs were information about: cancer, late effects and rehabilitation services and support from peers and professionals to cope with the new situation. Younger, caregiver comorbidity and recurrence of cancer of the patient were associated with more needs. Conclusions The results indicate that caregivers of cancer patients participating at the MBC report moderate caregiver burden, however, numerous caregivers reported need for support within several areas. Our findings need to be confirmed in a larger unselected group.


Introduction
Cancer and cancer treatment may have a major influence on the cancer patient, but also on their close relatives [1]. Adverse side effects are often most prominent during and the first weeks after treatment, but some last for months or even years after treatment and other might appear later (i.e., late-effects) [2,3]. As cancer becomes a chronic disease for some patients, symptoms of cancer and side effects might be lifelong [4]. Over the last few decades, there has been an increasing number of patients receiving treatment at outpatient clinics. Thus, a larger number of patients are dealing with symptoms of cancer and adverse side effects of treatment at home [5,6]. Due to less inpatient care, close relatives might have faced a greater responsibility the last years.
To the best of our knowledge, no studies have explored the burden and needs among caregivers of patients participating in rehabilitation or educational programs. The Montebello Center (MBC) is a Norwegian resource center for coping with cancer. For the last 25 years, cancer patients have participated in one-week educational programs at the MBC. If the patient has a close relative, they are invited to participate as well. By examining the burden and needs of caregivers for cancer patients participating in such program we aim to expand the current knowledge about the burden and needs of caregivers of cancer patients referred to an educational program by their physician. Such knowledge is important to develop tailored content in educational programs for caregivers in a rehabilitation setting.
The aims of this study were therefore to examine the perceived caregiver burden among caregivers of cancer patients with various cancer diagnoses attending an educational program, and to explore the characteristics of caregivers in risk of high caregiving burden. Further aims were to examine the caregivers need for support and identify associated factors.

Design and participants
The study has a cross-sectional design and was carried out between May 2021 and February 2022. All cancer patients scheduled to attend a one-week inpatient educational program, in this time period, at the Norwegian Resource Center for Coping with Cancer (MBC) received by post an invitationpackage with written information about the study, a consent form and a questionnaire for their potential caregiver ahead of the program, and were asked to hand it further on to their closest caregiver (only one) if they had one. Consequently, patients without close caregivers also received the invitation. On arrival at the MBC, the cancer patients were asked to remind their caregivers about the questionnaire if they had not yet responded. In addition and at a later date, a reminder was sent by post with a pre-paid return envelope enclosed to the cancer patients that had attended the program and where no response from the potential caregiver was registered.
Inclusion criteria for the caregivers were age !18 years (checked by the response in the questionnaire) and able to understand the Norwegian language.

Assessments
Caregiver burden was assessed with the Caregivers Reaction Assessment (CRA) [8]. CRA consists of 24 questions scored on a 5-point Likert scale from 1 (strongly disagree) to 5 (strongly agree), covering five domains evaluating the caregivers' reactions to caring at the time of survey. The five domains included in CRA are; Caregiver esteem (7 items measuring the perceived positive aspects of caregiving), Impact on health (4 items referring to the caregivers' perception of how his/her health has suffered as a result of the obligations of caregiving), Impact on schedule (5 items indicating the perceived effort and difficulty of obtaining healthcare needs and making care-related arrangements), Impact on finances (3 items measuring economic costs and losses likely caused by caregiving), and Lack of family support (5 items assessing caregivers' sense of how other family members have left him/her to provide all of the patients' care). A mean sum score was calculated for each domain (sum score divided by number of questions in the specific domain), reflecting the unweighted mean item scores ranging from 1.0 to 5.0. Higher subscale scores indicate higher levels of burden, except for caregiver esteem where a high score indicates a low burden. The CRA has a well-established validity and reliability in caregivers of cancer patients in different stages of the disease [8,24,25]. The Norwegian version of the CRA has shown satisfactory psychometric properties and some concurrent validity has been confirmed [24].
In this study, a score of !3 was considered as 'high caregiver burden' for Impact of health, Impact of schedule, Impact on finances and Lack of family support, and for Caregiver esteem a score of <4 was considered as 'high caregiver burden', these cut offs are suggested in previous studies by Sandstedt et al. [26] and Van Hof et al. [16]. The cut offs used are explorative cut offs, because the choice of cut offs are not argued for or explained in those two studies, and the population tested in Sandstedt et al. [26] are caregivers to patients with various diseases [16,26]. For example, the following responses on the Impact on schedule domain would give a score of 4.6. 'I agree' (score 4) on item 1 ('My activities revolve around caring for _'), 'I totally agree' (score 5) on item 2 ('I have to stop in the middle of my activities to care for _'), 'I agree' (score 4) on item 3 ('I visit my family and friends less since I started taking care of _'), 'I totally agree' (score 5) on item 4 ('I have failed to fulfill my commitments since I started taking care of _') and 'I totally agree' (score 5) on item 5 ('It is difficult to find time to rest because of the constant interruptions').
Need for support was assessed by a modified 13-item questionnaire developed by The Norwegian Cancer Society (Supplementary material, Additional file 1) [27]. Users have been part of the development of themes for the questions, and the questions have been pilot-tested for clarity and understandability [27]. A variable was constructed by adding the number of needs the caregiver reported (response range: 0-13, continuous variable). Thus, a higher number represents a need for more support. The variable is only a count of reported needs and not weighted for severity or urge for each item.
Background variables were self-reported and included the caregivers sociodemographic and medical variables; gender, age, education ( 13 years or >13 years), work situation (work full/part time or retired or sick leave/disability benefits/other), comorbidities were assessed through one question: have you any long lasting [>12 months] physical and/or mental illness, injury or condition which lead to reduced daily function? (yes or no), and variables related to the cancer patients; relationship to the cancer patients (spouse/cohabitant/partner or child/parent/friend/other), cancer type, time since cancer diagnosis (months), and recurrence of cancer (yes or no).

Statistics
Descriptive statistics are presented by frequencies, proportions, mean values, and standard deviations (SDs) (Tables 1 and 2, Figures 1 and 2). The associations with CRA five subscale scores (dependent variables) and selected covariates: age, gender, education, caregiver comorbidity, and time since diagnosis and recurrence of cancer (for the cancer patient) were explored in linear regression analyses (Table 3). These covariates were selected based on earlier research and clinical considerations. All covariates were included and retained in the multivariate analyses. The same procedure was used to explore the associations with number of needs (response range 0-13) (dependent variable) ( Table 4). The results are presented as regression coefficients (B) with 95% confidence intervals (CIs) and p values. A p value <0.05 was considered to be statistically significant. The data set also included caregivers with missing data on one or more items and therefore n varies. No substitutions were done for missing items for CRA. The analyses were performed using SPSS version 28.0 for Windows (SPSS, Chicago, IL, USA).

Participants
A total of 464 cancer patients received the information and invitation to the potential caregiver and were asked to invite a close caregiver. A total of 185 (40%) caregivers consented and completed the questionnaire. The median age of the participants was 58.0 (range 18-81) years, 58% were male, 64% were highly educated, and 57% worked full or part time. Of all, 83% were spouse/cohabitant/partner to the cancer patient, 35% caregiver to a breast cancer patient, 31% caregiver to a patient with recurrence of cancer, and median months since diagnosis was 24.4 months ( Table 1).  Table 2).

Level of caregiver burden
Proportions of caregivers with high caregiver burden in Impact on schedule was 39%, Caregiver esteem was 28%, Impact on finances was 19%, Impact on health was 15%, and Lack of family support was 15% ( Figure 1).

Factors associated with level of caregiver burden
Multivariable analyses showed that younger age, caregiver comorbidity and being a female caregiver (borderline significant, p ¼ 0.05) were associated with a higher caregiver burden on Impact on health. For Impact on schedule, younger age, higher education, caregiver comorbidity, recurrence for the cancer patient, and shorter time since diagnosis were associated with a higher caregiver burden. Younger age and caregiver comorbidity were associated with a higher burden on Impact on finances. Higher burden in Lack of family support was associated with younger age. No associations were found with a higher burden in caregiver esteem (Table 3).

Need for support and associated factors
The three most needed areas for information and support were as follows: 77% of the caregivers reported current need for information on cancer and late effects, 74% reported current need for information on cancer rehabilitation services, and 64% reported current need for support from peers ( Figure 2). Each caregiver reported in median 4.0 needs (range 0-13) (data not shown). Multivariate analyses showed that factors significantly   Numbers included varies because of missing data. a Higher score indicates a higher burden of caregiving, except for caregiver esteem, where a low score indicates higher burden. SD: standard deviation associated with more needs were younger age, caregiver comorbidity, and recurrence for the cancer patient (Table 4).

Discussion
The overall findings from this study reveal that caregivers of cancer patients participating at the MBC report moderate caregiver burden. Level of caregiver burden was highest in the area Impact on schedule. A proportion from 15% to 39% reported high caregiver burden in the different domains. The results demonstrated that being a female caregiver, younger age, higher education level, having caregiver comorbidities, recurrence in the cancer patient, and shorter time since diagnosis were associated with higher risk of caregiver burden in different domains. More than half of the caregivers reported need for information on cancer and late effects, cancer rehabilitation  and support from peers or professionals to cope with the new life situation. Younger age, comorbidity, and recurrence in the cancer patient were associated with more needs.

Level of burden
In agreement with other studies using CRA to measure caregiver burden in different phases of the cancer trajectory, our study showed similar levels in all five domains [13][14][15][16][17]20,25,28]. In line with the findings of this study, Impact on schedule was the domain in which burden was greatest in other studies [13,15,16,20]. The score of Impact on schedule in our study (mean score 2.6) was similar to studies of caregivers of advanced and palliative treated patients (mean scores 2.7-2.9) [13,17], to two other studies with a majority of caregivers of curative  [14,20] and in a longitudinal study (mean score 2.6 at baseline) [16]. Compared to the results in a recent study among caregivers of head and neck cancer patients shortly after diagnosis [16], we found a lower proportion of caregivers with high caregiver burden in all domains except for Caregiver esteem. One explanation could be the time of the survey and that only one diagnosis was investigated, in contrast to this study.

Factors associated with level of caregiver burden
In accordance with the findings of the studies by Astrup et al. [20], Utne et al. [13], Jansen et al. [18], and Van Hof et al. [16], our study also indicated that female caregivers experienced higher burden in Impact on health compared to male caregivers. In agreement with our study, Astrup et al. also found that those with comorbidity reported higher burden in Impact on health than those without comorbidity. The additive struggle of comorbidity in the caregiver concurring with taking care of a cancer patient was also confirmed by Maguire et al. [15] among caregivers of colorectal cancer patients mainly 1-2 years after diagnosis. In a review, association between younger age and higher caregiver burden was stated [19]. This is consistent with our findings, where caregivers with younger age reported higher caregiver burden in Impact on health. Younger caregivers may have other responsibilities, such as raising a family or building a career compared to older caregivers, and these factors could be stressful and have a negative impact on health and global quality of life [16].
In our study, those younger reported higher burden in Impact on schedule. Individuals in the beginning of midlife are often very busy including work and career, children and activities, social events, etc., might be a reasonable explanation for these findings. We also found that caregivers with higher education reported higher burden in Impact on schedule, probably explained by the difficulties of combining work and career with time-consuming care. In agreement with our findings, a recent study among caregivers of head and neck cancer patients found an association between higher education levels and higher caregiver burden in Impact on schedule [16]. These findings might be because caregivers from higher socioeconomic classes have a type of work that implies more responsibility, pressure to deliver, and a busy schedule both in their work and leisure time. The association found in our study between shorter time since diagnosis and increased burden on Impact on schedule is in line with the finding from Van Hof et al. [16] that during the long-term follow-up caregiver burden decreased in Impact on schedule. One explanation to this could be that during the first phase of the cancer trajectory there are more frequent hospital visits and treatment appointments etc. Caring for a cancer patient with tumor Stage III and IV were found to be associated with higher burden on Impact on schedule [16], and that could be seen in accordance with our findings that higher caregiver burden on Impact on schedule was associated with recurrence for the cancer patient.
Younger age and caregiver comorbidity were associated with Impact on finance, supporting that being ill and having health issues, especially in both caregiver and cancer patient, could be stressful for the economy. The association between younger age and higher burden in Impact on finance was also found among caregivers of colorectal cancer patients [15] , and one explanation could be the fact that an established economy is often achieved later in life. Lack of family support was associated with younger age. This highlights our findings that those of younger age have a higher risk of caregiver burden compared to older caregivers in four of five domains. We found no associations between the explanatory factors and caregiver esteem. The reasons for this could be the limited use of the caregiver esteem scale since the reported range used is only from 3 to 5 and none of the caregivers reported the highest-burden score 1 or 2 in this scale, or it could be other explanatory factors that are important to caregiver esteem that are not included in this study.

Need for support and associated factors
In line with the study of Sklenarova et al. [29] among caregivers of advanced cancer patients, caregivers in our study reported need for information as the most reported need. In a recent study, need for information was the second highest need among caregivers [30]. Also summed up in systematic reviews, the need for information on cancer, treatment and care was the most commonly reported [21,23]. Comparable Table 4. Factors associated with increasing need for information and support (range 0-13) in linear regression analyses. to our findings, other studies have also reported need of support among caregivers in the following domains: information on side effects, social welfare, and psychosocial adaptions [29]. In agreement with our findings, a qualitative study from Denmark among caregivers of cancer patients treated in an outpatient setting found that caregivers' need for support included support from peers and talking to healthcare professionals [31]. A previous study from the MBC by our group showed that 51% of the female caregivers and 28% of male caregivers reported need for support from peers [32], indicating more caregivers are reporting needs for that type of support in this study. Younger age, caregiver comorbidity and recurrence of the cancer for the patient were associated with more need for support. In compliance to our findings, a recent study found that comprehensive needs were higher among younger caregivers and those who cared for patients with recurrence [30]. We found similar variables associated to a greater number of needs and to higher caregiver burden that could help identifying vulnerable risk groups that should receive special attention early in the cancer trajectory in order to receive target care.

Strengths and limitations
A strength of our study is use of a validated and established instrument for caregiver burden (CRA). Limitations of the study are several. First, the participation rate was modest, but it seems that this rate vary from 25 to 90% in other studies including caregivers [33]. It is a limitation that we neither determined (1) how many of the participating cancer patients that had a close caregiver, nor (2) how many cancer patients with a close caregiver that actually passed on the information and invitation to their caregiver to answer a questionnaire. To recruit caregivers to studies could be challenging because it depends to some extent on the relationship between the patient and the caregiver. The modest response rate leave concerns of non-response bias risk. We do not know if the non-responders differ significantly from the responders. However, our results of caregiver burden and need for support may be under-reported, given caregivers with a high burden and high need are less inclined to answer the questionnaire due to lack of energy. On the other hand, it could also be that those with higher burden and more need for support are more likely to participate because it is of relevance for them.
Second, the study was based on a convenience sample and a formal power calculation was not performed. Caregivers invited were those nominated by cancer patients about to attend an educational program which is a selected group [34], therefore, our sample may not be representative of all those caring for cancer patients. An overrepresentation of highly educated among the cancer patients who attend the MBC [34] compared to the 37% with higher education in the Norwegian population [35] and among caregivers that responded to the questionnaire, is a concern in terms of generalization. Third, since the caregivers were nominated and invited by the cancer patients, we had no information concerning the non-responders.
Fourth, the lack of testing for validity and reliability on the need for support questions used in the study. Unfortunately, such validated measures translated to Norwegian do currently not exist as far as we know and the present measures used are based on face validity. The constructed variable that added the number of needs as an overall load of support needs was created based on all need items being equally weighted. It is possible that one severe need in one of the areas would be more problematic for the caregiver compared to several, but less severe, needs within different areas. However, in order to construct a variable with appropriate severity weighting, a level of specific information beyond the extent of this study would be required. Consequently, there is limited psychometric evidence for using the scale as explored in this study. Fifth, even though we found significant findings in caregiver burden between subgroups, we cannot state that these differences are clinically relevant, due to the fact that at present no minimal clinically important differences for the CRA are available [36]. In addition, there is a limitation that the CRA cutoffs used were not standardized. Lastly, since we performed several statistical comparisons there is a risk of statistical significant findings occurring by chance.

Ethical considerations
The study is a quality improvement study, evaluated and approved by the Privacy and Data Protection Officer at Oslo University Hospital (ePhorte number 21/05105). The study was presented for the South-East Regional Committee for Medical and Health Research Ethics (ref. number 204292) that considered the study outside their mandate. Written informed consent was provided from all participants. Regarding the recruitment procedure such as the caregivers being invited by way of the cancer patients and also including patients who possible did not have a close caregiver could be improved in future studies.

Clinical implications
A comprehensive cancer care model involves meeting the needs of caregivers [37]. Investigating the caregiver burden, associated factors, and need for support among caregivers of cancer patients is an important step in the development of targeted services. Knowledge from this study can be of help when facilitating rehabilitation servicessimilar to the MBC directed to caregivers to prevent impaired health and strengthen the caregivers' ability to face psychological and social challenges. The future content in the educational programs for caregivers at the MBC should focus on information on cancer and late effects, and cancer rehabilitation services, as well as include support from peers for example in group discussions and lectures on how to cope with the new life situation.

Conclusion
In conclusion, the results indicate that caregivers of cancer patients attending an educational program report moderate caregiver burden. The burden was greatest on caregiver schedule, and efforts to alleviate the burden on this domain should be emphasized. Being a female, younger age, highly educated, caregiver comorbidity, recurrence of cancer in the cancer patient and time since diagnosis were all associated with higher caregiver burden. More than half of the caregivers reported need for support within several areas, and associated factors to more needs were younger age, caregiver comorbidity, and recurrence for the cancer patients. It is important to identify subgroups of caregivers in high risk of burden stemming from a need of support, in order to maximize potential support with available resources. Our findings need to be confirmed in a larger unselected group preferably with the use of a validated measure concerning need for support.