This study evaluates the quality of life (QOL) and mental health (MH) of caregivers of patients with advanced cancer who are receiving ambulatory oncology care and associations with patient, caregiver and care-related characteristics.
This study evaluates the quality of life (QOL) and mental health (MH) of caregivers of patients with advanced cancer who are receiving ambulatory oncology care and associations with patient, caregiver and care-related characteristics.
Patients with advanced gastrointestinal, genitourinary, breast, lung or gynaecologic cancer, and their caregivers, were recruited from 24 medical oncology clinics for a cluster-randomized trial of early palliative care. Caregivers completed the Caregiver QOL—Cancer scale and the Medical Outcomes Study Short Form, version 2, and a questionnaire including care-related factors such as hours/day providing care and change in work situation. Patients completed a demographic questionnaire and measures of their QOL and symptom severity. Associations of these factors with caregiver QOL and MH were examined using linear regression analyses.
Of the 191 caregivers, 84% were spouses/partners, 90% cohabited with the patient, half were working and 25% had a change in work situation since the patient's diagnosis. On multiple regression analysis, better caregiver QOL was associated with better caregiver MH and patient physical well-being and with not providing care for other dependents. Worse caregiver MH was associated with female caregiver sex, worse patient emotional well-being, more hours spent caregiving and change in the caregiver's work situation.
Caregivers of ambulatory patients with advanced cancer may have compromised QOL and MH associated with worse patient physical and emotional well-being and with simultaneously caring for others and working outside the home. Early palliative care interventions directed at patient symptoms and caregiver support may improve QOL in this population. Copyright © 2011 John Wiley & Sons, Ltd.
In the last few decades, cancer care has shifted from the hospital inpatient setting to ambulatory settings. Combined with an ageing population and increased fiscal pressure on health care systems, this shift has resulted in families and other informal caregivers assuming increasing responsibility for patient care . Patients with advanced cancer often receive cancer treatment well into the palliative phase of their illness, with referrals to hospice or palliative care occurring late in the disease course [2-4]. The caregiving requirements are considerable for these seriously ill outpatients with cancer, who face complex medical problems as well as toxicities of treatment, without the support of a hospice team. However, few studies have examined the quality of life (QOL) of this group of caregivers . It is important to do so in order to evaluate whether and how these caregivers might benefit from early palliative care interventions.
Caregiver QOL is a multidimensional construct encompassing a broad range of elements including physical health, emotional well-being, social functioning, financial welfare and spirituality . It is related to, but distinct from caregiver burden , a construct originating in the geriatric literature, which has been defined as the perception of distress related to caregiving tasks . Caregiver QOL is by definition subjective and may vary between individuals based on differing personal values, perceptions and the stage of disease in patients for whom care is being provided .
A recent review of randomized trials for caregivers of cancer patients identified a need for research to determine patient and caregiver factors that could be targets for intervention . Further, a review of research on the QOL of caregivers of cancer survivors revealed few studies beyond the acute survivorship phase of diagnosis and treatment . Despite the heightened caregiving burden in the palliative phase of illness , studies of QOL at this stage are particularly sparse. Those that have been conducted were in mixed samples, comparing QOL in patients with advanced disease to those in acute phases of illness , or were small studies of patients who were already receiving hospice care [12-14]. To our knowledge, only one study has focused specifically on the QOL of caregivers of patients in the ‘late palliative phase’, which the authors described as ‘a condition beyond the curative phase, but [which] is not considered to be terminal’ ; these patients had metastatic cancer but a clinical prognosis of more than 4 months. This study, which was conducted in Norway, compared caregiver physical health and mental health (MH), measured by the Medical Outcomes Study Short Form (SF-36), with previously reported gender-specific norms. Physical QOL was higher than the norm in both genders, whereas MH was lower in male caregivers. This study did not, however, assess factors associated with QOL and MH in this population.
It has been theorized that variables associated with the caregiving experience and MH of caregivers of patients with cancer can be categorized into patient, caregiver and care-related factors . We have used this model to categorize possible factors associated with caregiver QOL. Potential patient-related factors include demographic and disease-related characteristics; caregiver-related factors include demographics, physical health and MH, other roles (occupational, family and social), living arrangements and the quality of the relationship with the patient; and care-related factors include the intensity and duration of care, help with caregiving and change in activities as a consequence of caregiving . Caregiver QOL has been found to be more strongly related to the MH than to the physical health of the caregiver [6, 16]. Therefore, in addition to examining the influence of MH on QOL, as has been done previously , we also examined factors that may specifically affect caregiver MH.
The purpose of the present study was to examine the QOL and MH of caregivers of ambulatory patients with advanced cancer and identify associations with patient, caregiver and care-related variables. We hypothesized that caregiver QOL would be related to caregiver characteristics, such as mental and physical health, female sex and younger age; to the QOL of the patient and to care-related characteristics, such as the amount of care provided and presence of and/or impact on roles other than that of caregiver. For caregiver MH, we had similar hypotheses but also that there would be an association with patient emotional well-being.
Patients and their caregivers were recruited from 24 outpatient medical oncology clinics between December 2006 and May 2010, at Princess Margaret Hospital, a comprehensive cancer centre in Toronto, Canada. This study was undertaken as part of a cluster-randomized controlled trial of early intervention by a specialized palliative care team versus conventional oncology care in patients with advanced cancer, the main purpose of which was to assess whether early palliative care improves patient QOL; caregiver QOL was included as a secondary outcome. The methods have been described previously  but are summarized below.
Eligible patients had a diagnosis of stage IV gastrointestinal, genitourinary, breast, or gynaecological cancer, stage III/IV lung cancer, or locally advanced pancreatic or oesophageal cancer. Other patient eligibility criteria included an Eastern Cooperative Oncology Group performance status score of 0–2  and a clinical prognosis of 6 months to 2 years (both determined by the patient's primary oncologist). Each patient was asked to identify his/her primary caregiver, who was subsequently approached in person (if present) or by telephone to request study participation and to seek informed consent. Exclusion criteria for both patients and caregivers were age less than 18 years or insufficient English to complete the questionnaires; patients were also excluded if they obtained a low score on a cognitive screening test  (Short Orientation–Memory–Concentration Test score <20, or >10 errors).
All caregivers and patients provided written informed consent; those who declined to proceed with the trial were asked for written consent to complete baseline measures only. Patients and caregivers completed measures at baseline and monthly for 4 months; for this study, baseline data were used for all analyses. The study was approved by the Research Ethics Board of the University Health Network.
Patients and caregivers completed a demographic questionnaire at accrual; items included age, gender, ethnicity, religion, level of education, marital status, living arrangement, employment status and household income. Other items included on this questionnaire were the caregiver's relationship to the patient; the number of hours per day spent providing care and whether caregivers had any chronic illness or disability, cared for others in addition to the patient, had formal or informal help caring for the patient or had experienced a change in their work or living situation following the patient's diagnosis. Patient medical records were reviewed by research staff to obtain patient cancer diagnosis and stage, cancer treatment status and comorbid diagnoses.
Caregiver QOL was measured using the Caregiver QOL Index—Cancer (CQOLC). This is a multidimensional tool developed through in-depth interviews with caregivers of patients with cancer and is the only validated measure designed specifically for caregivers of patients with cancer . It has been internationally validated, including in curative and palliative settings [16, 21], and uses a five-point Likert-type scale to measure physical, emotional, social family/financial and spiritual elements of caregiver QOL. Each item is rated from 0 to 4; the 35 items are summed to generate the total score.
The Medical Outcomes Study Short Form, version 2 (SF-36v2), was used to measure caregiver health and functioning . Its 36 items are organized into eight subscales: physical functioning, role limitation due to physical health problems, bodily pain, general health, vitality (energy/fatigue), social functioning, role limitation due to emotional problems and MH. Items from these subscales are combined into two summary scores: physical component summary and mental component summary, evaluating physical health and MH, respectively . These summary scores were calculated using SF-36v2 scoring software, which uses norm-based scoring: scores are T-transformed so that in a US norm population, the mean is 50 and the standard deviation 10. It is well validated [23, 24] and has been used as an outcome in other studies assessing the health of caregivers of cancer patients [10, 25].
The Functional Assessment of Cancer Therapy—General is a 27-item internationally validated measure of health-related QOL  and was used in this study to measure patient QOL. It has four subscales measuring physical, social/family, emotional and functional well-being, which are summed to obtain the total Functional Assessment of Cancer Therapy—General score.
The Edmonton Symptom Assessment System is a validated, self-administered tool, which uses a numerical scale from 0 to 10 to measure the severity of nine main symptoms in patients with advanced disease (pain, fatigue, drowsiness, nausea, anxiety, depression, appetite, dyspnoea and sense of well-being) . Because no time window is stipulated, we added instructions that symptoms were to be rated based on the previous 24-h period . The Edmonton Symptom Assessment System Distress Score is the sum of the nine symptom ratings and ranges from 0 to 90, with higher scores indicating greater symptom burden .
Additional patient measures included the Eastern Cooperative Oncology Group measure of performance status  and the Charlson Comorbidity Index, which generates a weighted score based on the presence of various medical illnesses  and is the most commonly used measure of comorbidity for patients with cancer .
Descriptive statistics were calculated for all variables. Simple linear regression was used to determine the impact of patient, caregiver and care-related characteristics on caregiver QOL and MH; backward stepwise multiple regression was used to determine factors that were independently associated with caregiver QOL. Variables significant at p ≤ 0.05 in the simple regression analyses were included in the multiple regression model. Employment status was excluded due to collinearity with age. When both patient and caregiver versions of a single variable were significant on simple linear regression analysis (e.g. age), only the caregiver variable was included. Analyses were performed using SPSS, version 19.0 (IBM Corporation, Somers, NY, USA). A two-sided p ≤ 0.05 was considered statistically significant.
Of the 1016 eligible patients approached for enrolment into the randomized controlled trial, 477 identified a caregiver; 105 caregivers were not approached for the following reasons: patient refused for the caregiver (n = 41), patient withdrew before the caregiver could be approached (n = 20), the caregiver could not be reached (n = 12) and reason not indicated (n = 32). Of the 372 caregivers approached, 182 consented and were enrolled; a further nine consented to completion of baseline measures but not to participation in the full trial. Thus, the study sample included 191 caregiver–patient pairs.
Caregiver and patient demographic and medical characteristics are shown in Table 1. The majority of caregivers were women, lived with the patient and/or were their spouse/partner; 35% reported caring for others, in addition to the patient (children ≤ 18, n = 44; parents, n = 13; spouse/partner, n = 7; other family member, n = 21). Caregivers spent a mean of 3.5 h/day caring for the patient (3.7 h/day for female and 3.3 h/day for male caregivers), and 40% (76/191) had formal (13/76) or informal (69/76) help. Informal help was from family (59/69), friends (4/69), family and friends (5/69) or unknown (1/69); formal help was through the home care system (11/13) or privately paid (2/13). Approximately half of caregivers were employed, and 25% described a change in work situation, the most common being a reduction in work hours (12% of participants; Table 2). Mean scores for caregiver and patient outcome measures are presented in Table 3.
|Characteristic||N (%)||Characteristic||N (%)|
|Age [median (min–max)]||61.0 (28–88)||Age [median (min–max)]||57.0 (22–83)|
|Female sex||89 (46.6)||Female sex||124 (64.9)|
|Primary tumour site||Relationship to patient|
|GI||72 (37.7)||Spouse/partner||160 (83.8)|
|Lung||31 (16.2)||Son/daughter||21 (11.0)|
|Genitourinary||34 (17.8)||Other||10 (5.2)|
|Gynaecology||21 (11.0)||Marital status|
|Breast||33 (17.3)||Married/common law||182 (95.8)|
|Marital status||Separated/divorced||1 (0.5)|
|Married/common law||176 (92.1)||Single||7 (3.7)|
|Separated/divorced||8 (4.2)||Living with patient||171 (89.5)|
|Single||1 (0.5)||Caring for others (other than patient)||67 (35.1)|
|Widowed||6 (3.1)||Non-European ethnicity||34 (17.8)|
|Living with others||184 (96.3)||Religion|
|Caring for others||55 (28.9)||Catholic/Christian||133 (69.6)|
|Non-European ethnicity||29 (15.2)||Non-Catholic/Non-Christian||13 (6.8)|
|Non-Catholic/Non-Christian||16 (9.0)||High school or less||65 (34.6)|
|Atheist/agnostic/none||37 (20.9)||College/university/other||123 (65.4)|
|High school or less||67 (35.4)||Retired||68 (35.6)|
|College/university/other||122 (64.6)||Employed||94 (49.2)|
|Employment status||Unemployed/Student||25 (13.1)|
|Retired||85 (44.5)||On disability||4 (2.1)|
|Unemployed||26 (13.6)||<$14 999||5 (3.6)|
|On disability||41 (21.5)||$15 000–29 999||13 (9.4)|
|Income||$30 000–59 999||36 (26.1)|
|<$14 999||6 (4.4)||≥$60 000||84 (60.9)|
|$15 000–29 999||13 (9.6)||Change in work situation||48 (25.1)|
|$30 000–59 999||38 (28.1)||Change in living situation||17 (8.9)|
|≥$60 000||78 (57.8)||Hours/day spent caregiving [mean (SD)]||3.5 (5.5)|
|Treatment status||Help caring for patient||76 (39.8)|
|Receiving active treatment||141 (73.8)||Own chronic illness or disability||35 (19.7)|
|Awaiting new line of treatment||28 (14.7)|
|Not on active treatment||22 (11.5)|
|Total sample, n = 191|
|Change in work situation||48||25.1|
|Less hours worked||22||11.5|
|On leave (four paid, four unpaid)||8||4.2|
|Scale and subscale(s)||Mean (SD)||Scale and subscale(s)||Mean (SD)|
|ECOG performance status||SF-36v2|
|0||52 (27.2)||Physical component summary (PCS)||52.5 (8.9)|
|1||124 (64.9)||Mental component summary (MCS)||41.0 (12.3)|
|Charlson Comorbidity Index [mean (SD)]||0.49 (0.8)||Caregiver QOL—Cancer||98.8 (15.8)|
|ESAS Symptom Distress Score||25.1 (15.4)|
|Physical well-being (PWB)||18.7 (5.7)|
|Social well-being (SWB)||22.8 (3.7)|
|Emotional well-being (EWB)||16.5 (5.0)|
|Functional well-being (FWB)||17.0 (6.1)|
Results of the simple and multiple regression analyses examining caregiver and patient factors related to caregiver QOL and MH are shown in Tables 4 and 5. On multiple regression analysis, better caregiver MH and better patient physical well-being were associated with better caregiver QOL, whereas providing care for additional individuals other than the patient was associated with worse caregiver QOL. Because items pertaining to MH are included in the CQOLC scale, we conducted another analysis in which we removed all items related to MH from the CQOLC. The SF-36 mental component summary score remained significantly associated with caregiver QOL (p < 0.0001).
|Variable||Simple regression analysisa||Multiple regression analysis (adjusted R2 = 0.456)|
|SF36-v2 mental component summary||0.817||0.1||<0.001||0.761||0.1||<0.001|
|Caring for others||−9.067||2.4||<0.001||−6.214||2.3||0.007|
|Change in work situation||−8.607||2.6||0.001|
|Income ≥ $60 000||5.919||2.8||0.038|
|ECOG performance status score||−5.432||2.1||0.010|
|Edmonton Symptom Assessment System|
|Physical well-being subscale score||0.700||0.2||0.001||0.565||0.2||0.004|
|Emotional well-being subscale score||0.763||0.2||0.001|
|Functional well-being subscale score||0.538||0.2||0.007|
|Variable||Simple regression analysisa||Multiple regression analysis (adjusted R2 = 0.193)|
|Spouse (relationship to patient)||5.320||2.4||0.030|
|Number of hours/day spent on caregiving tasks||−0.382||0.2||0.024||−0.389||0.2||0.010|
|Change in work situation||−8.090||2.0||<0.001||−8.356||2.0||<0.001|
|Receiving no cancer therapy||−6.180||2.8||0.027|
|Edmonton Symptom Assessment System|
|Physical well-being subscale score||0.439||0.2||0.007|
|Emotional well-being subscale score||0.497||0.2||0.006||0.712||0.2||<0.001|
|Functional well-being subscale score||0.367||0.2||0.017|
Factors associated with worse caregiver MH on multiple regression analyses were female caregiver sex, change in work situation, spending a larger number of hours per day on caregiving tasks and caring for patients with worse emotional well-being.
To our knowledge, this study is the first to assess factors associated with the QOL and MH of caregivers of ambulatory patients with advanced cancer who are not yet receiving support from formal hospice or palliative care services. Caregivers were generally in good physical health, as indicated by the SF-36v2 physical component summary score of 52.5, which is higher than the US population norm of 50 and higher than that reported in caregivers in a home hospice setting (46.0). However, the score for the mental component summary was only 41.0. This is not only lower than the US norm of 50 but also lower than the values of 44.4 in men and 43.5 in women recorded in a Norwegian sample of caregivers of ambulatory patients with advanced cancer . It is also lower than the value of 44.0 found in a study of caregivers in a US home hospice setting , despite the fact that overall caregiver QOL in our sample was not as compromised as that of caregivers in that study (CQOLC score 99 vs 85). Thus, there appears to be considerable mental strain associated with caring for patients with advanced cancer who are still receiving hospital-based ambulatory care.
Despite the fact that the caregivers in this study were caring for patients with relatively good performance status, caregiver QOL correlated strongly with patient physical well-being. Similarly, caregiver MH was associated with patient emotional well-being. There are few studies that have compared QOL of caregivers to that of patients . A small study of 30 ovarian cancer patient–caregiver pairs showed a direct relationship between the QOL of women with cancer and the QOL of their caregivers , as did two studies in cancer survivors [11, 32]. Our findings are consistent with these studies, as well as with others showing that the QOL of family members is worse in those caring for patients with advanced disease [33, 34] and that there is an association between caregiver and patient depression scores . In contrast, a study in the hospice setting found no association of caregiver QOL with the patient's physical state . This may be due to the support provided by health care professionals in hospice care or due to the more uniformly poor patient QOL in this population.
There are few studies documenting the socio-economic hardship experienced by caregivers of patients with cancer. In the SUPPORT study, where family members of seriously ill patients hospitalized in the USA were interviewed post-discharge, a family member had to quit work or make another major life change to care for the patient in 20% of cases . In the current study, 25% of caregivers underwent a change in work situation (the majority working fewer hours, with a minority quitting their job, taking leave or losing employment) after becoming a caregiver for the patient. This is a substantial proportion, considering that close to half of the caregivers in our sample were retired, and the patients being cared for had a relatively good performance status, were receiving ambulatory cancer care and had a clinical prognosis of greater than 6 months. Changes in work situation and number of hours spent caregiving were independently associated with worse caregiver mental well-being. Although, to our knowledge, the association between changes in work situation and caregiver QOL has not previously been reported, studies in Turkey  and Norway  have found that being unemployed was associated with worse caregiver QOL. In another study, increased emotional distress was reported by caregivers with lifestyle interference and limitation in their ability to participate in valued activities and interests, regardless of how much care they were providing . It is difficult in all of these studies, including our own, to interpret the directionality of the association between reducing work or other activities and worse QOL. However, it is reasonable to conclude that one goal of interventions for caregivers of outpatients with advanced cancer should be to provide some relief from long caregiving hours and to enable continued engagement in meaningful activities or employment if this is desired.
Independent of female sex, QOL was also worse for caregivers who had other caregiving responsibilities in addition to those for the patient. These findings are similar to those of a study in cancer survivors , and emblematic of a demographic and cultural circumstance, in which caregivers, who are most often women, provide care for multiple dependents of different generations . Our finding that young age was strongly associated with poor QOL and MH on simple linear regression analysis, as in some other studies [11, 41], but not in multiple regression analyses, suggests that the distress of those who are younger could be related to the role strain of caring for multiple dependents  or to the strains associated with changes in work situation. In this vein, a previous study found that employed caregivers of cancer survivors, who were also taking care of children, reported greater levels of caregiver stress .
Consistent with other studies [6, 44, 45], caregivers' mental, but not physical health, was strongly associated with caregiver QOL. Female sex was associated with worse MH, consistent with other studies of cancer caregivers [34, 37, 46]. Possible explanations include that women might perceive themselves to have less choice in the assumption of a caregiving role or set higher standards for themselves as caregivers. In one study, male caregivers were more likely than female caregivers to judge the caregiving experience as boosting their self-esteem ; others have theorized that greater perceived negative aspects of caregiving are associated with decreased MH [47, 48]. There may also be differences in the caregiving tasks assumed: female caregivers tend to take on more personal care and time-consuming tasks than male caregivers [49, 50], and the latter may receive greater recognition for caregiving and less often have multiple dependents to care for.
Strengths of our study include that we assessed an understudied population, used a QOL measure specific for caregivers and included patient and care-related factors in our analysis. Limitations include that the patient and caregiver population was generally well educated, of relatively high income and of predominantly European ethnicity. Similar to other studies of cancer caregivers [6, 11, 38, 43], we included patients with a variety of different cancers. There might be patient-related factors that affect caregiver QOL, which are specific to caring for patients with particular cancer diagnoses, although tumour site group was not associated with caregiver QOL in this sample. The cross-sectional nature of the data allows only for the description of associations between variables and not for definite statements about causality. Although the percentage of variance in caregiver QOL related to variables in the multiple regression model was approximately 46%, it was lower at 19% for the MH model. Our study did not include contextual factors, such as quality of the caregiver–patient relationship and attachment style, which have been shown to influence caregiver MH in other studies [15, 51]. Longitudinal studies are needed in this population, as is research on other factors that may influence caregiver MH, and with caregivers of patients who have specific cancer diagnoses.
In summary, we have evaluated factors predicting caregiver QOL and MH among caregivers of outpatients with advanced cancer. Although such caregivers are caring for patients who are less ill than those in hospice populations and may provide fewer hours of actual hands-on care, they also have less access to hospice or palliative care services and are more likely to have other commitments in addition to their caregiving responsibilities. Our results indicate that these caregivers experience substantial distress, particularly related to their MH, and that interventions at this early palliative stage of illness are indicated for caregivers as well as patients. The relationships of caregiver QOL with patient physical well-being, and caregiver MH with patient emotional well-being, indicate that early palliative care interventions at the level of the patient might also benefit the caregiver. Also, the influence on caregiver QOL and MH of care-related factors, such as caring for other dependents, the amount of care provided and change in work situation, supports a role for interventions providing social assistance for caregivers, even at this relatively early stage of advanced cancer.
We extend our thanks to the staff of the medical oncology clinics for their facilitation of this research. This research was funded in part by the Canadian Cancer Society (grant nos. 017257 and 020509; C. Zimmermann) and by the Ontario Ministry of Health and Long Term Care. The views expressed do not necessarily reflect those of the funding agencies, which also had no role in the study design, data collection, analysis and interpretation, writing or decision to submit for publication. This study was previously presented in part at the American Society for Clinical Oncology Annual Meeting, 2011, Chicago, IL, USA, and at the Canadian Association for Psychosocial Oncology Conference, 2011, Toronto, Canada.