| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
| ||||||||||||||||||||||||||||||||
a Department of Social Medicine, Academic Medical Center, University of Amsterdam, The Netherlands
b Northern Center for Health Care Research, University of Groningen, The Netherlands
c Department of Clinical Psychology, University of Groningen, The Netherlands
d Department for Health Services Research, National Institute of Public Health and the Environment, RIVM, Bilthoven, The Netherlands
Correspondence: Chris Nijboer, PhD, National Institute of Health Promotion and Disease Prevention (NiGZ), DeBleek 13, Postbus 500, 3440 AM Woerden, The Netherlands. E-mail: cnijboer{at}nigz.nl.
Decision Editor: Vernon L. Greene, PhD
 |
Abstract |
|---|
 TOP Abstract MethodsResultsDiscussionReferences |
|---|
Key Words: Caregiver burden • Cancer • Health • Longitudinal patterns
Many types of cancer at the present time could be described as chronic diseases requiring long-term treatment and posing numerous demands on patients and their primary caregivers. Alterations in the patient's functional ability, bodily functions, appearance, employment status, family and social role, and self-image are reported to have a direct impact on the caregiver (
Siegel, Raveis, Houts, and Mor 1991
;
Stetz 1987). With earlier diagnoses, advances in treatment, longer survival times, and a trend toward outpatient treatment, informal caregivers are likely to be involved in providing more complex care for a longer period of time. Research in the field of caregiving has mainly concerned family members of patients with mental disorders such as schizophrenia or dementia (
Baumgarten 1989;
Schulz, O'Brien, Bookwala, and Fleissner 1995;
Zarit and Toseland 1989). To date, however, only a few studies have specifically examined caregiver outcomes over time in the case of cancer. This study is an attempt to contribute to a more detailed insight into outcomes of caregivers of cancer patients over time.
One of the concepts that seems to play a central role within the caregiver situation is "caregiver burden," which refers to negative feelings that may be experienced when giving care (e.g.,
Vitaliano, Young, and Russo 1991). Studies have consistently reported that partners, in comparison to other informal caregivers, are more likely to experience caregiver burden (
Cantor 1983), become ill themselves (
George and Gwyther 1986), and experience higher rates of psychiatric symptoms than other types of caregivers (
Schulz, Visintainer, and Williamson 1990). Because partners are less likely than other informal caregivers to receive assistance (
Horowitz 1985), they may be identified as the most vulnerable group of caregivers. More recent studies have stressed that experiences of caregiving should be studied as a multidimensional construct, referring to physical, psychological, financial, and/or social experiences to specific care demands (
Given et al. 1992;
Siegel et al. 1991). And, although caregiving is mainly addressed as being burdensome because of its unpleasant consequences, caregiving may also be perceived positively (
Given et al. 1992;
Kinney and Stephens 1989;
Lawton, Kleban, Moss, Rovine, and Glicksman 1989;
Motenko 1989). Therefore, the most balanced approach to measure caregiver experiences incorporates both the negative and positive aspects of giving care.
Also, in cancer caregiver research, caregiver experiences are defined over a broad range of varying constructs that can be regarded as positive, neutral, and/or negative; subsequently, overall conclusions from studies are hampered by inconsistent findings. One of the first few longitudinal studies involving cancer patients and their spouses, conducted by
Oberst and James 1985, revealed that the intensity of partner's distress did not change over time, whereas the amount of distress reported by patients systematically decreased over time.
Keitel, Zevon, Rounds, Petrelli, and Karakousis 1990 reported similar patterns of distress in a study assessing the impact of the surgical treatment of cancer on the spouse. On the other hand, it was demonstrated that caregivers experienced high levels of burden when they first occupied the caregiver role, but these experiences dissipated with time (
Blood, Simpson, Dineen, Kauffman, and Raimondi 1994). Also, it was observed that caregivers continued to report similar levels of cancer caregiver's impact on financial problems, lack of family support, and loss of physical strength over time, but a decrease in impact on schedule over time and an increase on one positive dimension of caregiver experiences, caregiver's self-esteem (
McCorkle et al. 1993). Evidently, these findings indicate a need to examine patterns of caregiver experiences more thoroughly.
Caregiving can be conceived as a dynamic, ongoing process for which there may be several trajectories (
McCorkle et al. 1993;
Nijboer et al. 1998;
Oberst and James 1985;
Schulz and Williamson 1991). Several researchers in the field of oncology stress that experiences of caregivers of cancer patients have to be examined in relation to the phase of illness the patient is going through (
Northouse and Stetz 1989). These phases have been divided into three stages: the initial or acute phase, the chronic phase, and the resolution (
Rait and Lederberg 1990). During the acute phase the family is shocked, stunned, and frightened. Patients and their family members may use this period to find ways to cope and deal with the crisis. In the chronic phase, when primary treatment has been endured and the patient has been dismissed from the hospital, family members have to take on new and additional responsibilities. During the third phase, resolution, the family anchors itself in either survivorship or the bereavement process. All three periods may lead to considerable anxiety and perceived pressure in family members, particularly in the primary caregivers (
Kurtz, Kurtz, Given, and Given 1997). In order to gain insight into caregiver experiences over time, researchers increasingly emphasize the importance of including the acute and early chronic phases of the illness in studying caregiver experiences over time (
Given and Given 1991;
McCorkle et al. 1993;
Northouse and Swain 1987).
Patterns over time can be examined by a number of different analyses, depending on the specific research question. These methods may distinguish themselves particularly with respect to the endpoint level of the outcome(s) under study, the overall rate of change over time, or overall changeability (
Deeg, Smit, and Beekman 1997). Preceding the question, Which factors influence change within the caregiving situation?, is the course of the pattern over time itself. In order to obtain a more specific insight into patterns of caregiver experiences, changes over time can be determined at two different levels, across different groups and within one person over time (
Mulder, Ranchor, Sanderman, Bouma, and Van den Heuvel 1998). As far as we know, longitudinal caregiving research has predominantly focused on overall variability (i.e., groups' variability), and no data are known about the proportion of cancer caregivers who perceive caregiving more negatively, more positively over time, or who remain stable.
Research into cancer caregiver experiences over time is rather scarce, but research into patterns among certain types of caregivers is even more sporadic. In cross-sectional studies, relatively consistent associations were reported on caregiver experiences and sociodemographic characteristics such as gender, age, and socioeconomic status (SES). Women, especially at a younger age and with a lower SES, tend to perceive caregiving as more negative than their counterparts (
Barusch and Spaid 1989;
Blood et al. 1994;
Oberst, Thomas, Gass, and Ward 1989;
Siegel et al. 1991;
Young and Kahana 1989; Todd, Zarit, & Zarit, 1986), even when this perceived impact is adjusted for the amount of care provided and/or the level of patient's health (
Horowitz 1985).
The aim of the current study is to describe overall patterns of caregiver experiences in partners of cancer patients over time; across different groups (i.e., according to gender, age, and SES), and within individuals over time. Data will reflect on the potential influence of the early and chronic phases of the disease (i.e., the 2 weeks prior to hospital admission, and to 3 and 6 months thereafter). Caregiver experiences are considered as a multidimensional construct, including negative and positive dimensions.
|
Methods |
|---|
|
TOPAbstractMethodsResultsDiscussionReferences |
|---|
In order to describe patterns of caregiver experiences over the first period of time, representing the acute and early chronic phases of the patient's illness, and consistent with former research, three measurement points were chosen: the time of patient's diagnosis (T0), and 3 months (T1) and 6 months (T2) thereafter. Baseline measurement (T0) took place shortly after hospital admission and within 2 weeks after patient's surgery, with questions being asked retrospectively, referring to the period before hospital admission (i.e., the period before the caregiver role was likely to be taken on). The second measurement point (T1) took place 3 months after baseline and the third measurement (T2) occurred again 3 months later (i.e., 6 months after baseline). Patients and partners were interviewed face-to-face by trained research assistants using structured questionnaires (e.g., CRA), and they also completed a self-report questionnaire.
Measurements
Patient's Descriptive Variables
In order to get information about potential selection bias, several descriptive variables of the patients were collected: patient's diagnosis, whether the patient had a stoma or not, cancer-related symptoms, comorbidity, and level of depression.
Cancer-related symptoms were measured using the Rotterdam Symptom Check List (RSCL;
De Haes, Knippenberg, and Neijt 1990). Patients were asked to rate the severity of a symptom, either a physical (e.g., nausea) or a psychological symptom (e.g., worrying), on a 4-point scale 3 months after diagnosis. The internal consistencies of the two scales, "physical symptoms" and "psychological symptoms," were 0.83 and 0.89, respectively.
Comorbidity was measured by asking the patient if he/she had a disorder of chronic nature in the preceding 3 months (e.g., stroke, serious heart problems, high blood pressure, diabetes). The short-form 16-item list of the Central Bureau of Statistics was used (Centraal Bureau voor Statistik, 1989).
The level of depression was measured by the 20-item Center for Epidemiological Studies Depression scale (CES-D;
Bouma, Ranchor, Sanderman, and van Sonderen 1995;
Radloff 1977). The CES-D comprises a 20-item self-report scale that taps the level of depressive symptoms during the week preceding hospital admission. Total scores can range from 0–60, with higher scores indicating a higher level of depressive symptoms. Cronbach's 
was 0.91.
Partner's Sociodemographic Variables
Age, gender, and SES were used to define subgroups in order to study changes in caregiver experiences across different groups. The three age groups on the basis of age at baseline were: <50 years, 50–65 years, and >65 years. SES was operationalized as the highest level of education attained, and was categorized into three levels: low (primary school only), middle (lower/general secondary schooling and intermediate vocational schooling), and high (higher vocational schooling and university).
Caregiver Experiences
Caregiver experiences were measured using the Caregiver Reaction Assessment scale (CRA), developed by
Given and colleagues 1992. The CRA consists of 24 items that contain four negative dimensions and one positive dimension: Disrupted Schedule, Financial Problems, Lack of Family Support, Loss of Physical Strength, and Self-esteem. The subscale Disrupted Schedule (5 items) measures the extent to which caregiving interrupts the usual activities of the caregiver. The subscale Financial Problems (3 items) measures the financial strain on the caregiver as a consequence of the caregiving situation. The subscale Lack of Family Support (4 items) assesses the extent to which the caregiver perceives a shortage of family support and the caregiver's perception of being abandoned in their caregiving responsibilities. The subscale Loss of Physical Strength (5 items) assesses the caregiver's feeling of deterioration in physical health. The subscale Caregiver's Self-esteem (7 items) aims to measure the extent to which caregiving contributes to individual self-esteem. Respondents were asked to rate the perceived impact of caregiving on a 5-point Likert scale. For each subscale, a total score was computed reflecting the mean item, with a range between 1.00 and 5.00. A higher score represented a greater amount of the attribute. The subscales of the CRA were found to be valid and reliable (Cronbach's -coefficients ranged from 0.68–0.90;
Given et al. 1992;
Nijboer, Triemstra, Tempelaar, Sanderman, and Van den Bos 1999).
Response
Of the 238 eligible cancer patients and their partners, informed consent was obtained from 181 at the onset of the study. The response rate was 76%. No response bias was found between participants and nonresponse with respect to the variables Age, Gender, Diagnosis, and Region. One hundred forty-eight (82%) couples participated in all three waves. Reasons for loss to follow-up were: serious illness (n = 15; 8%), refusal to report for follow-up (n = 10; 6%) and death of the patient (n = 8; 4%).
Patients who remained in the study did not differ regarding age, gender, diagnosis, comorbidity, and feelings of depression from dropouts. However, the proportion of patients with a stoma was slightly higher (p < .10), and they also reported more physical symptoms at baseline (p < .05) as compared to participants who were followed up to 6 months. Partners who remained in the study did not differ from dropouts regarding age, gender, and the baseline values of caregiver experiences: Disrupted Schedule, Financial Problems, Lack of Family Support, and Self-esteem. However, participating partners reported a significantly less negative impact on loss of physical strength at baseline, compared to those who dropped out over time (participants: mean score = 1.94; dropouts: mean score = 2.22; p < .05).
Statistical Analyses
Analyses were performed to examine changes in the five specific caregiver experiences domains over a 6-month period, starting from the time of diagnosis, referring to 2 weeks prior to patient's hospital admission (i.e., baseline, T0). Changes over time at group level were examined by means of General Linear Model (GLM) repeated measurement analyses of variances (ANOVAs) for the total group and across the separate subgroups. Post-hoc Scheffé tests for multiple comparisons of means were performed to gain insight into the actual differences between several groups. In addition, to indicate the relevance of the size of the difference between the average scores, Cohen's effect size statistic d for paired observations was calculated (
Cohen 1977). An effect size <0.20 indicates a small effect, 0.20–0.50 a moderate effect, and >0.80 a large effect (
Cohen 1992). Within each subgroup, caregiver experiences of the three waves were compared by means of paired t tests. The following comparisons were performed: wave T0 versus T1, T1 versus T2, and T0 versus T2.
Changes over time at individual level were assessed by means of calculating Pearson's correlation coefficients. Based on previous research concerning "change at an individual level" (
Sanderman and Ranchor 1994), the following rules for change were defined: no change at individual level was defined as a correlation coefficient of >0.70, coefficients between 0.70 and 0.50 indicated moderate change, and coefficients <0.50 were classified as highly dynamic.
Finally, the proportion and direction of changes in caregiver experiences were specified. The differences of the individual mean scores between the measurement points were calculated and subsequently classified as decrease, equal, or increase. For the four negative subscales, an increase indicated a more negative impact on the construct over time, whereas for the positive subscale Caregiver's Self-esteem, an increase indicated that more self-esteem was derived from caregiving. Again, there are no standard rules available to evaluate the size of a change as relevant. In order to correct for variability by chance (i.e., variability resulting from a slightly different answer), "change" was defined as a difference of at least half the baseline standard deviation of the subscale score between the two mean subscale scores (i.e., >0.5 x SD baseline value).
|
Results |
|---|
|
TOPAbstractMethodsResultsDiscussionReferences |
|---|
|
|
|
Individual Variabilities and Proportional Changes Over Time
In Table 3 , the variability at an individual level and proportional changes over time are presented for each subscale. Because changes between T0 and T1, and between T1 and T2 all pointed in the same direction, but were less substantial as compared to overall changes between T0 and T2, only changes between T0 and T2 are presented. As shown, the correlation coefficients of the subscales Disrupted Schedule and Lack of Family Support were relatively low, indicating a large individual variability. For Disrupted Schedule, variability was high within each subgroup. The frequency distribution of part B showed that most changes in the impact on disrupted schedule turned out to be in a less negative direction, indicating less negative impact on disrupted schedule over time. With respect to Lack of Family Support, individual caregivers reported a change in scores over time, even though no group variability was found (see Table 2 ). A closer inspection of the data revealed that the proportions of caregivers who reported a decreased, an equal, or an increased score 6 months after baseline were similar, and therefore the impact on Lack of Family Support remained stable at a group level. Particularly within the highest SES group, individual variability was large, and of all subgroups, those in the youngest age groups remained most stable at the individual level.
|
In sum, the largest individual variability was observed for the impact on Disrupted Schedule and Lack of Family Support, and the least individual variability was observed for the impact on Self-esteem. Within subgroups, individual variability depended predominantly on the type of impact. However, most individual variability was observed among caregivers in the lowest SES group.
|
Discussion |
|---|
|
TOPAbstractMethodsResultsDiscussionReferences |
|---|
Female caregivers perceived a more negative impact on Loss of Physical Strength as compared to male caregivers, and although Caregivers' Self-esteem did not change over time in male caregivers, female caregivers reported a less positive influence on self-esteem over time. Differences in social roles, in range of competing roles, and in role commitment may account for the observed distinctive patterns between female and male caregivers (
Kramer and Kipnis 1995). Also, differences in emotional attentiveness, in coping styles, or in occupying caregiving tasks may explain the observed gender differences (
Rose-Rego, Strauss, and Smyth 1998). As female caregivers were found to be more attentive to their emotions, they were more likely to report negative effects of caregiving. Also, women have been found to use more emotion-focused coping strategies, which are associated with higher reported levels of negative health outcomes (
Miller and Cafasso 1992). Moreover, female caregivers traditionally are likely to perform more personal care and household chores, which may be more time-consuming, demanding, and ongoing than traditional male tasks (
Rose-Rego et al. 1998). Our findings give support to the importance of making a basic distinction between female and male caregivers, rather than to analyze them as a homogeneous group of caregivers, but additional research is needed to interpret the observed gender differences in caregiving.
Younger caregivers tended to report a more negative impact on disrupted schedule and continued to do so over time, whereas older caregivers appeared to experience caregiving as less negative over time. A possible explanation may be that younger persons feel more restricted in performing outside activities, such as work and social activities, when they have to take on the caregiving role. Likewise, younger persons may experience more secondary role strains, such as work and child rearing; thus, caregiving may be a more easily accepted role at a later stage of life.
Regarding SES, no differences between the separate categories were found with respect to the negative caregiver experiences. However, scores for the impact on self-esteem in caregivers with a low SES were high and remained practically constant over time, whereas caregivers with a high SES derived substantially less and a decreasing amount of self-esteem from caregiving. As caregivers with a low SES derive self-esteem from caregiving activities—and they remain to do so over time—the caregiver role may be of greater importance to them compared to those with a high SES. It is plausible that caregivers with a high SES derive more self-esteem from outside activities (e.g., work) or from other resources.
Patterns of three out of five caregiver experiences were found to be stable. Findings of stable patterns of caregiver experiences have also been reported by other researchers (
Collins, Stommel, Wang, and Given 1994;
Pruchno, Kleban, Michaels, and Dempsey 1990). Hence, the moment at which someone takes on and perceives the role of caregiver is of great importance. In the current study, baseline has been defined as the period of 2 weeks before hospital admission. However, someone may have assumed the caregiver role long before this period, perhaps at the time of the first complaints. The implication to study patterns of caregiver experiences among carers of cancer patients at the moment of the first complaints (i.e., in most cases before the diagnosis is actually clear) is, however, unfeasible. In this study, the acute and early chronic phases of the cancer process were addressed. The fixed time period of 6 months after patient's discharge might be too short to test, however, if patterns in caregiver experiences have become manifest. Patterns in caregiver experiences may even become more pronounced over time, when the patient's health is still deteriorating or improving and with ongoing changes in care demands or in caregivers' health (
Kurtz, Kurtz, Given, and Given 1995). Also, it might well be that the longer the chronic phase, the more caregivers' resources get depleted, and this may have an effect on how caregiving is perceived. As the consequences over a longer term are still largely unknown, future research should reflect on a longer period of time.
In order to explain the patterns observed, additional longitudinal analyses should be performed (
Singer 1998). The focus of this study was purely descriptive, and therefore implies reservation regarding the predictive value of time (i.e., whether a wear-and-tear model or an adaptational model is most appropriate). However, the current findings gave little support to using only the wear-and-tear model, that is, a model which simply suggests that the longer care is provided, the more negative (or less positive) the outcomes will be, and which is supported by previous research (
Townsend, Noelker, Deimling, and Bass 1989;
Walker, Acock, Bowman, and Li 1996). Caregivers were shown to derive less self-esteem from caregiving activities over time, and this is in favor of the wear-and-tear model, but, on the other hand, the impact on disrupted schedule showed a decrease over time and this is consistent with an adaptational model. Apart from time and care characteristics, disease-specific characteristics, social support, personality characteristics, quality of the relationship, and caregiver's own health are likely to play an important role in caregiving (
Biegel, Milligan, Putham, and Song 1994;
Bull 1990;
Intrieri and Rapp 1994;
Kurtz et al. 1995;
McCorkle et al. 1993;
Nijboer, Tempelaar, Triemstra, Sanderman, and Van den Bos 1999;
Walker et al. 1996;
Wright, Clipp, and George 1993).
Some issues need to be addressed with respect to the generalizability of the findings. First, only caregivers of cancer patients with relatively good health and a good prognosis were included and stayed in the study. Over the three waves, 13% of the patients died or had to withdraw because of deteriorating health, and especially in their partners the consequences of caregiving might have become very obvious. Second, a relatively high score on the positive domain of caregiver experiences and relatively low scores on the negative domains were observed as compared to reference studies involving caregivers of cancer patients (
Given, Given, Helms, Stommel and DeVoss 1997;
Given et al. 1993;
Kurtz, Given, Kurtz, and Given 1994;
Kurtz, et al. 1995,
Kurtz, et al. 1997;
McCorkle et al. 1993). It may be questioned whether high scores are due to a ceiling effect, and low scores to a floor effect, respectively. Furthermore, a general problem in the discussion on change scores is the phenomenon of "regression to the mean." In addition to studying caregiver experiences in other circumstances, the sensitivity of the CRA instrument needs to be further examined. Finally, the definition of the concept of change is rather arbitrarily defined, and one may wonder if the reported variability reflects actual changes or accidental fluctuations. Unfortunately, as yet, data concerning absolute changes in caregiver experiences as well as cutoff scores to detect caregivers at risk are unavailable. An external criterion is needed that is acceptable as a standard against which the instrument under study can be judged. The development of indicators for assessing clinically relevant changes of caregiving experiences should be encouraged.
In sum, this study shows the advantages of studying caregiver experiences as a multidimensional concept that includes both negative and positive experiences of caregiving and underlines the value of studying both inter- and intraindividual patterns across different subgroups. Women, younger caregivers, and those with a high SES were identified as subgroups that are most vulnerable as caregivers and perhaps in potential need of support. Further longitudinal research is needed to document the patterns of caregiver experiences over a longer period of time and for different groups of caregivers. Also, more research could reveal factors that differentiate caregivers who experience wear and tear from those who seem rather to adapt to the caregiving situation in order to develop specific interventions to support informal caregivers.
|
Acknowledgments |
|---|
We thank the patients and partners who took part in the study, and we gratefully acknowledge the physicians and coworkers in the cooperating hospitals for the recruitment of patients.
Received for publication March 18, 1999. Accepted for publication June 7, 2000.
|
References |
|---|
|
TOPAbstractMethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  J. N. Wells, C. S. Cagle, P. Bradley, and D. M. Barnes Voices of Mexican American Caregivers for Family Members With Cancer: On Becoming Stronger J Transcult Nurs, July 1, 2008; 19(3): 223 - 233. [Abstract] [PDF]
|
|
|
|
 |
|
 N. O'rourke Reliability Generalization of Responses by Care Providers to the Center for Epidemiologic Studies-Depression Scale Educational and Psychological Measurement, December 1, 2004; 64(6): 973 - 990. [Abstract] [PDF]
|
|
|
|
 |
|
 C. E. Jacobi, B. van den Berg, H. C. Boshuizen, I. Rupp, H. J. Dinant, and G. A. M. van den Bos Dimension-specific burden of caregiving among partners of rheumatoid arthritis patients Rheumatology, October 1, 2003; 42(10): 1226 - 1233. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||
| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
