sion tasks performed was predictive of all the mental health outcomes except one. When disturbance level was examined, giving emotional support in difficult ...
Research in Nursing & Health, 1993, 16, 251-263
Caregiving Tasks as Predictors of Mental Health of Wife Caregivers of Men with Chronic Obstructive Pulmonary Disease Sylvie Cossette and Louise Levesque
The type, number, and amount of disturbance generated by caregiving tasks as well as the adequacy of informal social support were examined as predictorsof the mental health of 89 wife caregiversof men with chronic obstructive pulmonary disease (COPD). The number of supervision tasks performed was predictive of all the mental health outcomes except one. When disturbance level was examined, giving emotional support in difficult situations was the task category that predicted five of the eight outcomes. Support variables accounted for small amounts of variance of most outcomes. However, caregivers who received respite support when they needed it were actually three times more likely to take psychotropicdrugs than those not receiving Support. 0 1993 John W h y & Sons, Inc
It is recognized that the mental health of family caregivers of frail elderly, demented, or physically dependent persons can be affected by the caregiving situation (for review see Schulz, Visintainer, & Williamson, 1990). However, although chronic pulmonary obstructive disease (COPD) is an important cause of disability among North American men aged 55 and over (Cooreman, Thom, & Higgins, 1990; Levasseur, 1983; Manfreta, Mao, & Litven, 1989), the mental health of caregivers of people with that disease has rarely been studied. Only Sexton and Munro (1985) have compared the mental health of wife caregivers of men with COPD with that of wives who were not caregivers; they found that life satisfaction was lower and stress higher among caregivers.
COPD affects neuropsychological functioning and can produce frequent somatic complaints, changes in mood, and a greater likelihood of feeling depressed and anxious (McSweeny, Grant, Heaton, Adams, & Timms, 1982; Prigatano, Wright, & Levin, 1984). The major clinical manifestations of COPD are shortness of breath and frequent exacerbation of the disease. Due to their breathing impairments, men with COPD find it difficult to express their feelings and to accomplish activities of daily living (ADLs) and instrumental activities of daily living (IADLs). If many of these men are still at home despite severe impairment, it is only because of their wives’ continuous investment in caregiving. This investment includes many tasks such as helping with ADLs, giving health care specific to COPD,
Sylvie Cossette, MSc, RN, is a professor at Facult6 des sciences infirmibres, Universite de Montreal. Louise Levesque, MSc, RN, is a professorat Facult6 des sciences infirmibres, Universite de Montreal and Centre hospitalier CBte-des-Neiges. This research was partly funded by the Canadian Pulmonary Association and by the Quebec Pulmonary Association. The authors express their appreciation to the staff of cooperating agencies for recruitment assistance, and to Francine Ducharme, PhD, RN; Louise Potvin, PhD; and Evelyn Adam, MN, RN; for their help in preparing the manuscript. This article was received on June 5, 1992, was revised, and accepted for publication March 11, 1993. Requests for reprints can be addressed to Sylvie Cossette, Facult6 des sciences infirmibres, Universite de Montreal, C.P. 6128, succursale A. Montreal (Quebec), Canada H3C 3J7.
0 1993 John Wiley & Sons, Inc. CCC 0160-6891/93/040251-13
251
252
RESEARCH IN NURSING 8 HEALTH
and providing emotional support, as well as taking on those responsibilities and household activities previously assumed by their husbands (Sexton & Munro, 1985). On the one hand, the types or categories of caregiving tasks are quite varied; on the other hand, it has been found that wife caregivers receive less help or support from their families than do husband caregivers (Enright, 1991; Pruchno & Resch, 1989). It has also been found that spouse caregivers, whether male or female, receive less help than other caregivers (BCland, 1991). Much remains to be learned about the relationship between extensive caregiving tasks, informal social support, and the mental health of caregiver wives of men with COPD. A better understanding of these relationships would be relevant in view of the fact that both caregiving tasks and social support can be modified by appropriate nursing interventions. The aim of this study was to evaluate the influence of various categories of caregiving tasks, as well as the influence of social support, on the mental health of wife caregivers of men with COPD. As noted in recent reviews (Biegel, Sales, & Schulz, 1991; Zarit, 1989), certain themes are common to many caregiving models, especially those based on a theory of stress (e.g., that of Lazarus & Folkman, 1984). In addition to the importance of social support and situational variables, such models emphasize the distinction between the objective demands of caregiving (e.g., illness severity, the number of behavioral impairments or the number of tasks performed) and the subjective evaluation of these demands (e.g., disturbance, that is troublesomeness or inconvenience, stressfulness, or burden) in predicting the mental health of caregivers. In a review of caregiving studies (Zarit, 1989), this distinction between objective demands and their subjective evaluation has been studied more often from the point of view of demented persons’ behavioral impairments than from that of caregiving tasks. The relationship between the mental health of caregivers and the caregiving demands was therefore variable, depending on the frequency and the number of the carereceiver’s behavioral impairments (the objective demands) and on the amount of disturbance or stress occasioned by these impairments (subjective evaluation of the demands). Concerning caregiving tasks and mental health, Zarit (1989) reports that very few investigators have made a distinction between objective demands and their subjective evaluation with regard to different categories of tasks. Biegel et al. (1 991) contended that “it would be useful to mea-
sure both the tasks carried out by caregivers to assist disabled relatives and the extent to which carrying out these tasks results in inconvenience and discomfort” (p. 52). The objective demands of caregiving include illness severity and caregiving tasks carried out by the caregivers. Biegel et al. (1991) found that, on the whole, the illness severity of the carereceiver was negatively but moderately related to caregiver well-being. The objective demands of caregiving tasks are often measured by scales that furnish composite scores based mainly on assistance in ADLs and IADLs (Albert, 1991) rather than being based on a wide variety of tasks (e.g., the scale of Oberst, Thomas, Gass, & Ward, 1989). Whatever tasks are included, the composite scores do not allow for assessment of those categories of tasks that are most relevant to the mental health of the caregivers. Composite scores measure tasks either in terms of their number, or in terms of the amount and frequency of assistance, or, alternatively, in terms of the number of hours involved in care. In many cases, significant links with mental health are observed. For instance, the total number of tasks is a significant predictor of objective burden (Montgomery, Gonyea, & Hooyman, 1985) and perceived burden (Jutras & Veilleux, 1991), while the number of hours spent in caregiving tasks is associated with depression (Baumgarten et al., 1992). Oberst et al. (1989) examined caregiver load, that is, the time and energy spent by family members in caregiving tasks. The outcomes considered were four evaluation dimensions as measured by the “Appraisal of Caregiving Scale” and the results differed according to the outcome examined. Caregiver load was correlated with only two dimensions: harm/loss and threat associated with caregiving. Comparing wife caregivers and husband caregivers, Pruchno and Resch (1989) found that the amount of assistance is a significant predictor of global burden only for the former; for neither gender was the amount of assistance related to burden (17-item scale) or to depression. Stommel, Given, and Given (1990) examined the frequency of assistance in relation to a sense of entrapment associated with caregiving, while Tennstedt, Cafferata, and Sullivan (1992) focused on the relationship between depression and the number of hours spent in performing tasks. Results from both studies indicated no significant links between objective tasks and mental health outcomes. The distinction between various categories of tasks has been taken into account in only a few studies. Montgomery et al. (1985) considered
CAREGIVING TASKS / COSSETTE AND LEVESQUE
both the total number of tasks and five categories of tasks (e.g., grooming/nursing care; walking/transportation; feeding/toileting; bed/wheelchair transfer; and telephone, money/medication management) in relation to objective and subjective burden. In the regression analysis for objective burden, only two of the five categories were significant (grooming/nursing care and walking/transportation), while none was significant as a predictor of subjective burden. The authors explained their results not in terms of the time required to perform the tasks but rather in terms of the lack of freedom resulting from the mere performing of these tasks. Jutras and Veilleux (1991) observed that caregivers who participated in personal care (ADLs) and those who interacted with professionals on behalf of the elderly experienced greater perceived burden than those who did not. These authors, as well as Horowitz (1982), underlined the direct commitment and intimacy involved in providing personal care. However, when the number of hours of care and the frequency of tasks were considered for tasks related to ADL and IADL, neither category predicted negative emotional reactions to the caregiving situation (Given, Stomme], Collins, King, & Given, 1990). Finally, Stoller and Pugliesi (1989) found that the number of hours of assistance per month had a different effect on mental health depending on whether ADL or IADL tasks were considered. Only a greater number of hours assisting with ADL increased burden and emotional distress, and there was no influence on caregiver depression. When the number of hours of assistance was considered without distinguishing the categories (ADL and IADL), that variable predicted more burden but less emotional distress and less depression. Furthermore, these same authors found that the total number of tasks, regardless of categories, was not a predictor of any of the three outcomes. Subjective evaluation of the demands on caregivers has also been examined. In one study (Carey, Oberst, McCubbin, & Hughes, 1991), the evaluation of task demands (labeled “caregiving burden”) was carried out by multiplying the time spent in 14 caregiving tasks (objective demands) by the amount of difficulty associated with their performance. Regardless of the category of task, caregiving burden had an indirect effect on mood dysfunction via negative appraisal of caregiving. Sheehan and Nuttall (1988) used a composite measure of objective demands (amount of assistance in ADLs and IADLs, labeled “percent of
253
care”) and of subjective reaction (amount of distress associated with the elder’s need for assistance, labeled “reaction to ADLs”). Reaction to ADLs was a significant predictor of the two outcomes examined, personal strain and negative emotions, while percent of care was not a significant predictor. However, percent of care was the only variable significantly correlated with three mental health indicators, depression, somatization and anxiety, while reaction to ADLs was significantly correlated only with anxiety. The correlation coefficient between percent of care and reaction to ADLs was only .08, an indication of the importance of the distinction between objective and subjective caregiving demands. Finally, Haley, Levine, Brown, and Bartolucci ( 1987) found a significant correlation between perceiving more stress and less self-efficacy in managing ADL and IADL tasks and caregiver depression. Those who reported less self-efficacy in IADL tasks also reported a lower level of life satisfaction. In terms of social support, as Lazarus and Folkman ( 1984) and House ( 198I ) point out, it seems that it is not necessarily the objective support received or available (e.g., frequency, amount, etc.) that is protective of mental health but rather the subjective appraisal of support (e.g., satisfaction, adequacy, etc.). According to a recent survey of literature on the social support of caregivers of demented persons (Ltvesque & Cossette, 1991), the links between social support and caregiver well-being are more often significant when evaluative measures are considered than when objective ones are used. Moreover, Gottlieb and Selby (1989) report gender differences concerning social support. They quoted Hirsch’s (1979) results indicating that even if the women in his study received more emotional support than men, they were less satisfied than men with the support they received. Haley et al. (1987) used an evaluative measure of informal social support, a composite measure of satisfaction with the network, and found that greater network satisfaction was significantly correlated with less depression and higher life satisfaction. The authors also used two objective measures of social support (the number of friends and of close relatives); the results indicated that more friends decreased only depression, while the number of close relatives was not related to either measure. Most of the investigators who have examined task variables have used objective measures of social support, either actually received or merely available, and their results are inconclusive. Ben-
254
RESEARCH IN NURSING B HEALTH
eficial influences of social support on objective burden (Montgomery et al., 1985) and on perceived burden (Jutras & Veilleux, 1991) were observed. Stommel et al. (1990) used three objective measures of social support and only one was a significant predictor of a sense of entrapment. Several authors found no significant relationship between objective measures of social support and mental health (Given et al., 1990; Pruchno & Resch, 1989; Stoller & Pugliesi, 1989; Tennstedt et al., 1992). Among situational variables, the age of the caregiver and of the carereceiver, the duration of the illness and of the caregiving situation, and the number of negative and positive life events in the last year may potentially influence the mental health of caregivers (Biegel et al., 1991; Murrell, Norris, & Hutchins, 1984). Other situational variables that may influence mental health include the diagnosis of the carereceiver, the gender of the caregiver, and the relationship between the caregiver and carereceiver (Dillehay & Sandys, 1990). To sum up, results from the foregoing studies underline the relevance of distinguishing various categories of caregiving tasks. The results also reveal that the mental health of caregivers has rarely been studied while concurrently taking into account various categories of caregiving tasks, as well as their number and the amount of disturbance caused by them. Furthermore, the adequacy of social support has seldom been examined in studies on caregiving tasks. On the basis of these observations, the following research questions were asked: Do the relationships between categories of caregiving tasks and the mental health of wife caregivers of COPD men differ according to whether the number of tasks is considered or whether the amount of disturbance occasioned by those tasks is considered? What is the influence of the adequacy of informal social support on the mental health of these wives?
METHOD Sample This cross-sectional study was conducted in Montreal (Canada) with a convenience sample of 89 French-speaking wife caregivers of men with COPD living at home. All men had to have a confirmed COPD diagnosis from a pneumologist and their illness severity, as estimated by the degree of impairment, had to be between moderately impaired (Grade 111) and severely impaired (Grade V) on a 5-point scale. Seventy-one sub-
jects were recruited through their COPD husbands who had come to four hospital respirology clinics in Montreal for a routine check up by their pneumologist. Such visits take place every 3 months for medication and pulmonary functioning check ups. Eighteen subjects were recruited through a chronic pulmonary home care service, before home care service had begun. An analysis of variance revealed no significant difference in illness severity between the two groups of COPD men. The mean for illness severity, that is the ratio of FEV, (forced expiratory volume in one second) to the normal FEV, was 36.18% and ranged from 13% to 60%. The mean duration of the illness was 13 years (SD = 14.83). The mean age of the COPD men was 68.6 years (SD = 7.03) while the wives’ mean age was 65 years (SD = 7.8 1). Only four caregiver/carereceiver dyads received formal services at the time of the study; they consisted of periodic home visits for dressing changes and were not related specifically to the COPD nor to the caregiver’s experience. In most cases, the wives felt they had been taking care of their husbands all their married life (M = 25 years) and had difficulty dissociating such habitual nurturing care from that specifically related to COPD.
Measures Mental health outcomes. Eight indicators of mental health were examined: six dimensions of psychopathology, level of stress, and psychotropic drug use. The SCL-90 scale. Six dimensions of the French version of the SCL-90 (Derogatis, Rickels, & Rock, 1976) were used (depression, obsession-compulsion, somatization, anxiety, hostility, and interpersonal sensitivity). Three other dimensions-phobic anxiety, paranoid ideation, and psychoticism-were omitted because a previous study (Lauzon, 1988) with spouse caregivers of persons with Alzheimer’s disease indicated a high percentage of zeros (no experience of the symptoms) on all items. The validation and normalization of the English version of SCL-90 were reported by Derogatis et al. (1976) and Derogatis, Meyer, and Diny (1981) and psychometric properties were satisfactory. Alpha coefficients for the retained six dimensions vary from .84 to .90 and the test-retest coefficients were between .78 and .86 (Derogatis, 1977). High convergent and discriminant validity between the SCL-90 and the MMPI (Minnesota Multiphasic Personality Inventory) also were observed (Derogatis et al., 1976).
CAREGIVING TASKS I COSSETTE AND LEVESQUE
The French version of the SCL-90 was validated and normalized for a Montreal population of women by Fortin, Coutu-Wakulczyk, and Engelsmann ( I 989); the alpha coefficients of the six dimensions retained for this study varied from .90 to .92. In the current sample, the alpha coefficients for five of the six dimensions were between .77 and .82, the exception being for hostility (alpha = .58). The Pearson correlation coefficients among the six dimensions were between .38 and .78, the latter being between anxiety and depression. The respondent was asked to indicate the degree of distress related to psychopathology symptoms on a 5-point scale ranging from not at all (0) to extremely (4). The score for each dimension was divided by the number of items therein. Higher scores on each of the dimensions indicate poorer mental health. The subjective Stress Scale. This scale (Chapman et al., 1966), which was used by Sexton and Munro (1985) with COPD wife caregivers, is composed of four statements; examples are: “In general, I am usually tense or nervous,” and “My daily activities are extremely trying and stressful.” The respondent answers how well each statement describes her. Answers range from not at all (1) to very well (4). Higher scores indicate higher stress. Croog and Fitzgerald (1978) reported a relative stability of the scale while Reeder, Schrama, and Dirken (1 973) observed significant positive correlation coefficients with social stress and neuroticism. In the present study the alpha coefficient was .82. The Pearson correlation coefficients between the stress score and the scores of the six dimensions of the SCL-90 varied between .43 and .69. Given these values, stress was considered as a distinct and complementary dimension of mental health. The use of psychotropic drugs. Psychotropic medication included prescribed antidepressants, anxiolytics, sedatives, and antipsychotics. Data regarding the use of drugs were obtained by the respondents’ self-report. Answers were dichotomized (yes = 1 or no = 0), since few wife caregivers use more than one kind of drug. Predictors. Caregiving tasks index. This index, composed of nine categories of caregiving tasks, was developed for the current study and takes into account the specific situation of the COPD wife caregiver. It is based on previous categorizations of tasks (Clark & Rakowski, 1983; Horowitz, 1982; Steinmetz & Amsden, 1983), on the Sickness Impact Profile questionnaire (Bergner, Bobbit, Carter, & Gilson, 1981; Bergner, Bobbit, Pollard, Martin, & Gilson, 1976), and on the clinical ex-
255
perience of the first author. Some of the items describe tasks similar to those included in measures used with various samples of caregivers, such as the Caregiving Load Scale (Oberst et al., 1989), the Beliefs about Caregiving Scale (Phillips, Rempusheski, & Morrison, 1989), and the Role Strain Scale (Archbold, Stewart, Greenlick, & Harvath, 1990). Content validity was examined by seven respiratory nurse specialists. Fifty-nine items divided among six categories were retained because they were classified in only one assigned category and were recognized by six of the seven experts as highly relevant to the caregiving situation of COPD patients. Thirty-nine of the 59 items were retained on the basis of principal component analysis (PCA) results. The items in each category explain between 52.2% and 63.6% of the variance of the factors. However, for three of the six task categories, PCA revealed two factors each and it was decided to consider the second factors independently, thus giving rise to three new task categories. The final version of the index has therefore nine categories: social interaction, which represents tasks associated with facilitating the husband’s social exchange with other people; sleep, which includes tasks associated with monitoring and helping with sleeping problems related to dyspnea or shortness of breath; supervision, signifying tasks associated with decision making or with warning the husband of dangerous behaviors; ‘‘light’’ emotional support, when faced with sadness, helplessness or complaints of discomfort; “difficult” emot ional support, when faced with fear of dying or argumentative behaviors; helping with ADLs, such as grooming; health care related to COPD; internal household management, such as routine household tasks; and external household management, which includes heavy tasks. Table 1 presents sample items for each category of tasks. Two indexes for each of the nine categories were computed from this questionnaire: the “number of tasks” index (sum of “yes” responses for each task done) and the “amount of disturbance” index. For the latter, subjects answered the following question each time they did a task: How much do you feel bothered, upset or affected by carrying out this task or this activity? The answers ranged from no? at all ( 0 ) to very much (4). The K-R 20 coefficients for the “number of tasks” indexes (dichotomous data) ranged between .51 and .83, except for the social interaction category (K-R 20 = .31). The Cronbach alpha coefficients for the “amount of disturbance” indexes ranged between .60 and .83. Cor-
256
RESEARCH IN NURSING 8 HEALTH
Table 1. Categories of Tasks and Sample Items Social interaction (7 items) -Go out with him to meet friends and family -Find for him some activity that he can do Sleep (3 items) -Watch him when he’s sleeping -Calm him in the night when he has difficulty breathing or cough Supervision (3items) -Help him to make decisions -Watch to prevent dangerous activity or activity he cannot do because of breathing problems Light emotional support (4 items) -Listen and encourage when he feels sad, worthless or depressed -Emotional support when he has pain or discomfort Difficult emotional support (5 items) -Listen and discuss when he talks about death -Tolerate aggressive behavior Internal housekeeping (4 items) --Do usual household work --Do laundry External housekeeping (3 items) -Do grocery shopping and errands -Money management (go to the bank, etc.) Grooming care (5 items) -Help with bathing -Help with dressing and undressing Health care (5 items) -Observe symptoms (cough, sputum, cyanosis, edema, pain) -Check and control humidity and aeration in house
relation coefficients among categories ranged between .09 and .53 when considering the number of tasks, and between -.05 and .72 for amount of disturbance. The adequacy of support by family and by friends. These two variables were measured using a 3-item index (Lauzon, 1988); each item corresponds to one function of social support (tangible assistance, emotional support, and respite). The family support index (alpha coefficient = .50) represents the wives’ evaluation of the frequency, from never (0) to most ofrhe time (3), with which the tangible assistance, the emotional support, and the respite given by their family met their needs. A sample item is “Are the members of your family giving you the help you need to take care of your husband?” The sum of the ordinal answers to the three items was used as a global score of the adequacy of family support. The friend support index (alpha coefficient = .61) is computed with the same three items but the subjects answered by referring to the support received from their friends. The Pearson correla-
tion coefficient between the two indexes is .13. Illness severity and situational variables. AS mentioned above, the illness severity of COPD husbands was estimated by the ratio of the FEV observed to the FEV, predicted normal, expressed as a percentage. Negative and positive life events during the past year were assessed through two open questions: answers from both questions were coded as one or more events ( I ) or no such events (0). Finally, the following socio-demographic variables were examined: husbands’ and wife caregivers’ age, duration of illness, and duration of caregiving situation as perceived by the wife. The duration of caregiving situation measure was rejected because the data were found to be unreliable. Procedure. Data concerning illness severity were collected from the patient’s hospital files during the pulmonary consultation. A consent form was signed by both the wife caregivers and the husbands. The interviews, which lasted approximately 2 hr, were conducted in the subjects’ homes. At no time was the subject retained if her husband’s pulmonary status had not remained stable since the check up in the hospital clinic.
,
RESULTS Descriptive data indicate that the mean scores are high on three out of the six dimensions of the SCL-90, somatization, obsession-compulsion, and depression (Table 2). Nearly 40% of the wives used psychotropic drugs. Taking into account the possible range of the scores for each of the categories, more tasks were carried out in the categories of external household management, internal housekeeping, and light emotional support than in the other categories. Difficult emotional support as well as light emotional support tasks occasioned more disturbance than other categories of tasks. The social support received from family was perceived as more adequate than that received from friends.
Number of Tasks Model and Disturbance Model For the seven continuous data outcomes (depression, obsession-compulsion, somatization, anxiety, hostility, interpersonal sensitivity, and stress) two hierarchical multiple regression analyses (number of tasks model and disturbance model) were estimated using SPSS-pc software (Norusis, 1990). The hierarchical procedure was used in order to isolate the contribution of control (situa-
257
CAREGIVING TASKS / COSSETTE AND LEVESQUE
Table 2. Means and Standards Deviations for Predictors and Dependent Variables Possible Range
Actual Range
M
SD
Number of Tasks
Variable Name
0-7 0-3 0-3 0-4 0-5 0-4 0-3 0-5 0-5
0-6 0-3 0-3 0-4 0-5 0-4 0-3 0-5 0-5
2.37 1.60 1.09 2.66 2.02 2.57 2.17 1.51 2.57
1.33 1.06 .98 1.39 1.61 1.05 1.07 1.52 1.52
Support adequacy Family support Friend support
0-9 0-9
0-9 0-9
4.58 1.57
3.24 2.14
.72 .70 .36 .63
.67 .61 .46 .62 .64 .37 3.65
0-4 0-4 0-4 0-4 0-4 0-4 4-1 6 (38.2%)
2.67 2.50 2.56 2.62 2.90 2.00 4-16
tional) variables and illness severity (step I), of social support (step 2), and caregiving tasks (step 3). To be retained in the final regression analysis, the control variables (including illness severity) in step 1 had to be significantly correlated with at least three outcomes. At that stage, illness severity, positive life events, carereceiver age and duration of sickness were dropped. The social support variables were retained in step 2 even if not correlated with mental outcomes in order to answer the second research question. In step 3, all categories of task were retained because each was a significant predictor of at least three outcomes in preliminary stepwise regression analyses. The number of tasks model (Table 3) revealed that, together, the variables explained a significant proportion of the variance of each of the seven mental health outcomes (overall R2 was between .20 and .36). Much of the variance was explained by the number of tasks variables (step 3), which contributed between 18% and 23% of the variance of six out of the seven outcomes, the exception being somatization. One task category (supervision tasks) was far more influential than
.58 .26 8.25
Actual Range
M
SD
Disturbance Related to Tasks
Categories of tasks Social interaction Sleep Supervision Light emotional support Difficult emotional support Internal housekeeping External housekeeping Grooming care Health care
Dependent variables Somatization Obsession-compulsion Interpersonal sensitivity Depression Anxiety Hostility Stress Psychotropic drugs use
Possible Range
0-18 0-12 0-9 0-27 0-15 0-16 0-9 0-15 0-20
0-12 0-9 0-6 0-12 0-1 5 0-1 0
0-8 0-1 1 0-1 1
1.06 .97 .48 3.23 3.49 1.18 .52 .45 1.19
2.25 1.77 1.27 3.77 3.98 2.21 1.22 1.46 2.14
the others, having a significant relationship with six out of seven outcomes, again with the exception of somatization. The number of internal household tasks predicted somatization, the number of light emotional support tasks predicted hostility, and the number of difficult emotional support tasks predicted obsession-compulsion. The adequacy of informal social support variables (step 2) had no significant impact on any outcome. The tasks disturbance model (Table 4) explained a significant proportion of the variance of five of the seven outcomes; R2 ranged from .26 to .40.The two outcomes that were not significantly explained were somatization and interpersonal sensitivity. The tasks disturbance variables explained between 15% and 25% of the variance of five out of seven outcomes, again with the exception of somatization and interpersonal sensitivity. One category of tasks, difficult emotional support, was particularly important as it significantly predicted somatization, depression, obsessioncompulsion, anxiety, and hostility. The disturbance generated by health care tasks was a signif-
258
RESEARCH IN NURSING & HEALTH
Table 3. Hierarchical Regression According to the Number of Caregivlng Tasks (Standardized Beta) Mental Health Outcomes ObsessionSomatization Depression Compulsion Anxiety Predictor Variables Step 1-Control Age Negative life events R2
Step 2-Support adequacy Family support Friend support R2 change Step 3-Number of tasks External house Internal house Light emotional support Difficult emotional support Supervision R2 change Total R2 F
Interpersonal Sensitivity Hostility Stress
Beta
Beta
Beta
Beta
Beta
Beta
Beta
-.16 .17 .07'
- .20'
-.12 .11 .06
-.26'* .12
-.28" .11 .07'
-.21' .26' .16'
-.21' .03 .06
- .06
-.17
- .08 .01 .01
- .04 - .01 .01
-.lo .oo
.oo
-.12 -.06 .02
.19' .12*
.08
.oo
.01
.03
'01 -.31** .07 .11 .09 .11 .20* 2.04'
.09 -.17 .02 .19 .32" .21 .36' 4.93'
icant predictor of depression. As in the number of tasks model, the contribution of the adequacy of social support (step 2) was negligible.
The Psychotropic Drugs Model Stepwise logistic regressions were performed to analyze factors associated with psychotropic drug use since that outcome variable had been coded dichotomously. To be entered into the final model, the control variables had to be significantly correlated with the outcome, and the task variables had to be significant predictors in preliminary stepwise logistic regressions including only the tasks variables. To control for the possible overlap between mental health and the use of psychotropic drugs, and based on discriminant analyses (standardized canonical coefficients), the somatization and depression scores were retained as predictors. The result was that four independent variables remained significant: number of social interaction and of supervision tasks, depression, and family support. At this point, in order to obtain the odds ratio values, which are easier to interpret than logistic coefficients (Clipp & George, 1990), for three of the four independent variables, their total scores were dichot-
.11 - .06
-.lo .30' .27' .20' .27* 3.11'
.lo'
.07 -.13 .16 .09 .34" .23*
.34' 4.34'
.02
.OO
.oo
-.lo - .08
-.11
- .03
- .22' .20 .37' .18' .35' 4.56'
- .06 .50" .19' .28' 3.36'
.08
.20 -.01 - .01 .11 .29' .18' .27' 3.14'
omized and forced into the final model. One particularity is the 3-item family support scale, which assesses on an ordinal scale the adequacy of three different types of support; dichotomization of the total score was not indicated because it would be difficult to interpret the results. It was decided to consider the three items separately, each one dichotomized as never received when needed (0) and received sometimes, ofien or most ofrhe rime when needed (I). The dichotomization cut off points are shown at the bottom of Table 5. The findings indicate that depression was the most important predictor of the use of psychotropic drugs (Table 5). The odds ratio value indicates that the wife caregivers who scored 60 or more on depression had 6.85 times the risk of using psychotropic drugs than had those who scored below 60. The performance of two or more tasks in one category also influenced drug use. The caregivers who carried out more than one supervision task had nearly three and a half times the risk of using these drugs than had those who did not. However, even though the value of the coefficient does not reach significance level, the wives who performed more than one social interaction task had about half the risk of using these drugs than had those who did not. The influence of family support differed according to
259
CAREGIVING TASKS I COSSETTE AND LEVESQUE
Table 4. Hierarchical Regression According to the Amount of Disturbance Related to Caregiving Tasks (Standardized Beta) Mental Health Outcomes ObsessionSomatization Depression Compulsion Anxiety Predictor Variables Step 1-Control Age Negative life events R2 Step 2-Support adequacy Family support Friend support R2 change Step 3-Task disturbance External house Internal house Light emotional support Difficult emotional support Health care Social interaction R2 change Total R2 F
Interpersonal Sensitivity Hostility Stress
Beta
Beta
Beta
Beta
Beta
- .01
-.06 .22' .08'
.02 .12 .04
-.13 .14 .04
-.13 .15 .04
-.01
- .07
.oo
- .02
- .01
.20 .06' - .02
-.11
.08
.oo
.01
.04
- .01 -.15 - .08 .26' .12
.09 .02 .03 34" 30'
-a
-
.06 .13 1.49
.08
.20* .40* 4.29'
- .02
Beta
- .09
.28" .12*
.03 .01
.oo .01
.02
.oo
.06 .09 -.11 .43" .27 - .08 .25' 30' 3.64'
.09 .10 .04 .26* .20 .02 .21 .26* 3.02'
.00
-.lo .13 -.14 37" .06 .03 .15' .27' 3.15*
.06 .01 .17 .28 -.12 .10 .16 1.62
Beta
- .09
.03 .01 -.12 - .04 .03
.oo .21 .04 .23 .26 - .06 .23' .27' 3.16'
a Social interaction task category was omitted in this equation because of statistical artifacts which are due to the multicollinearity between that variable and somatization and the difficult emotional support task category.
'p
9
.05.
*' p 5 .01.
Table 5. The Use of Psychotropic Drugs: Logistic Regression Psychotropic Drugs Use
Predictor Variables Mental health variablesa Depression Task care variablesb Number of social interaction tasks Number of supervision tasks Family support variablesc Instrumental assistance Emotional support Respite
Logistic Coefficients
Odds Ratio
1.92'
6.85
- .56
.57
1.23'
3.42
.12 - .01 1.12
1.13 .99 3.07
Mental health variables are dichotomized 0 = 59 or less and 1 60 or more on standardized score on the dimensions of the SCL-90. b Number of tasks variables were dichotomized 0 = none or one task done in the category and 1 = two or more tasks in the category. c Family support variables were dichotomized 0 = never received when needed, and 1 = received sometimes, often or most of the time when needed. * p I.05. a
=
the function of that support. The adequacy of instrumental assistance and of emotional support did not seem to influence the use of psychotropic drugs. However, subjects who felt that the respite given by their family met their need sometimes, o f e n , and most of the time were three times more likely to take psychotropic drugs than those who felt that such respite never met their need.
DISCUSSION The answer to the first research question is that the relationships between categories of tasks and the mental health effectively differ according to whether the number of tasks or whether the amount of disturbance is considered. In other words, the relationships are different depending on whether the objective demands or their subjective evaluation are taken into account. More specifically, two factors stand out as particularly related to mental health: the number of supervision tasks and the amount of disturbance associated with difficult emotional support. The answer to
260
RESEARCH IN NURSING & HEALTH
the second research question is that the part of family support known as “respite” actually seems to increase the probability of using psychotropic drugs. The regression analysis indicates quite clearly that the number of supervision tasks is a significant predictor of seven out of eight mental health variables, including the use of psychotropic drugs. This can be explained by the confining and continuous nature of those tasks. Hour after hour, day after day, the caregivers must be on the alert to the tasks of “warning the husband of dangerous behaviors” in order to maintain his psychological and physical security. This result is consistent with the observations of Montgomery et al. (1985), who found that caregiving tasks that are of a confining nature are related to greater burden. In the Sexton and Munro study (1985), a problem identified by wife caregivers was their loss of freedom. This loss of freedom is well described by Mace and Rabins ( 1 98 1) in their book The 36-hour day, addressed to caregivers of persons with Alzheimer’s disease; the authors point out that the need for constant supervision causes the caregivers to feel trapped and this in turn may affect their well-being. The amount of disturbance generated by difficult emotional support may be attributed to a lack of the necessary skills; a wife caregiver may simply not know how to respond to her husband’s distress. She may also come to believe that her husband’s demanding and argumentative behaviors are voluntary and manipulative when, in fact, they may be secondary to COPD (Sexton & Munro, 1985). In this study, as well as in the one by Carey et al. (1991), emotional support tasks caused the caregivers the greatest amount of disturbance. The amount of disturbance related to health care tasks was a significant factor in explaining depression. The wives may view their husbands’ health care requirements as a sign of the gravity of his condition and therefore fear their impending death (Sexton & Munro, 1985). As mentioned by Zarit, Orr, and Zarit (1985), the gradual decline of the carereceiver is routinely reported as causing feelings of depression in caregivers of demented persons. As well, the health care required by COPD men, for example, the monitoring of symptoms, may require skills and competencies that the wives do not feel they possess. In such a situation, the self-esteem associated with feeling efficient as a caregiver may be lowered, giving rise to depressive moods (Haley et al., 1987). However, the health care category of tasks that predicted depression in the current study dif-
fers from the categories found by Haley et al. (1987). In their study, the caregivers who felt more stressed and less self-efficient in ADL and IADL tasks reported higher depression. In the current study, the personal care related to ADLs and the internal or external household management (IADLs) were not linked to caregivers’ depression. Our results may differ from those of Haley et al. (1987) simply because their study did not have a health care task category. In the present study, the format of the question unfortunately does not reveal the specific reason for the amount of disturbance occasioned by health care tasks. The use of psychotropic drugs was predicted mainly by depression, which was considered as predictor in the model. The predictive influence of depression might have been enhanced by the fact that the depression score was dichotomized at the cut off point of light clinical symptoms (Derogatis et al., 1976). With respect to task categories, the number of supervision tasks was a significant predictor of the use of psychotropic drugs. Thus, the number of supervision tasks was the predictor related to the greatest number of mental health outcomes. Among the informal social support variables, only the adequacy of respite family support, even if measured by a single item, increased the probability of psychotropic drug use. However, this relationship was not in the anticipated direction and reached only a .10 level of significance. It may be that more respite was offered to those wife caregivers who, because of their use of psychotropic drugs, seemed to be in poor mental health. In the present study, the percentage of wife caregivers using psychotropic drugs is 38.2%. That percentage is higher than the one observed (30%) by Clipp and George (1990) for caregivers of demented persons. However, the current sample is older than Clipp and George’s sample and includes only women, rather than men and women. According to Lapierre and Adams’ study (1989), being female and older increases the prevalence of psychotropic drug use. Indeed, in their study, the sample of women over 65 from the general population of Quebec showed a 20.5% use of psychotropic drugs. While the subjects in the current study had a mean age of 65, their greater use of psychotropic drugs (38.2%)may perhaps be explained, at least in part, by their caregiving tasks, and not only by their age and gender. The caregiving tasks questionnaire has been used only in this study and probably needs further validation. The low reliability of the social sup-
CAREGIVING TASKS i COSSETTE AND L~VESQUE
port questionnaire also calls for a cautious interpretation of the results. The cross-sectional design does not permit conclusions about direction of causality and the nonrandom sample precludes generalization of the results. However, the sample used in this study, unlike many studies of caregivers, is composed of caregivers not recruited through a support group and it is homogeneous with regard to the type of relationship with the carereceiver and to the latter’s gender and illness. A theory of stress such as the one proposed by Lazarus and Folkman (1984) appears relevant to the study of relationships between caregiving tasks and the mental health of wife caregivers; it should guide further caregiving studies. Indeed, such a theory leads to a consideration of the distinction between objective caregiving task demands and the evaluation of those demands. Oberst et al. (1989) have already pointed out the importance of assessing the distressing or bothersome aspects of caregiving tasks. Findings from the current study indicate the importance of also distinguishing among various types of caregiving tasks because the relationships between tasks and mental health varied according to the categories. From a clinical point of view, studying caregiving tasks according to their type, number, and amount of disturbance occasioned appears to be worth considering, because various aspects of those tasks call for different interventions. For example, when it comes to the number of tasks, one strategy might be to mobilize and to strengthen informal supports as well as offering formal services in order to decrease the number of caregiving tasks done by the caregivers even though the wives may not ask for informal or formal services. As a case in point, none of the wives in the current study received formal services related to the COPD caregiving situation. The wives may have felt insecure about having others care for their husbands or they may have feared his blame or have worried that he would refuse to be left with someone else. They may also have been unaware of the advantages of such services. When the focus is on the amount of disturbance related to tasks, one strategy might be to encourage the caregivers to freely express their negative feelings, thereby reducing their impact. Guidance, emotional support, and information might also help caregivers to be better prepared to cope with different caregiving tasks, particularly dificult emotional support and health care. It is hoped that this research may shed some light on the difficult experience of wife caregivers of COPD husbands. Further research
261
must, of course, be undertaken before the complexity of the caregiving tasks may be fully understood and the appropriate interventions clearly identified.
REFERENCES Albert, S.M. (1991). Cognition of caregiving tasks: Multidimensional scaling of the caregiver task domain. The Gerontologist, 31, 726-734. Archbold, P.G., Stewart, B.J., Greenlick, M.R., & Harvath, T. (1990). Mutuality and preparedness as predictors of caregiver role strain. Research in Nursing & Health, 13, 375-384.
Baumgarten, M., Battista, R.N., Infante-Rivard, C., Hanley, J.A., Becker, R., & Gauthier, S . (1992). The psychological and physical health of family members caring for an elderly person with dementia. Journal of Clinical Epidemiology, 45, 61-70. Btland, F. (1991, November). The principal caregiver as the pivot in the relationship between sources of help. Paper presented at the 44th Annual Scientific Meeting of the Gerontological Society of America, San Francisco, CA. Bergner, M., Bobbit, R., Carter, W., & Gilson, B. ( 1 98 I). The Sickness Impact Profile: Development and final revision of a health status measure. Medical Care, 19, 787-805. Bergner, M., Bobbit, R., Pollard, W., Martin, D., & Gilson, B. (1976). The Sickness Impact Profile: Validation of a health status measure. Medical Care, 14, 57-67. Biegel. D.E., Sales, E., & Schulz, R. (1991). Family caregiving in chronic illness. Newbury Park: Sage. Carey, P.J., Oberst, M.T., McCubbin, M.A., & Hughes, S.H. (1991). Appraisal and caregiving burden in family members caring for patients receiving chemotherapy. Oncology Nursing Forum, 18, 134 I1348. Chapman, J.M., Reeder, L.G., Massey, F.J., Borun, R., Picken, B., Browning, G.G., Coulson, A.H., & Zimmerman, D.H. (1966). Relationship of stress, tranquilizers and serum cholesterol levels in a sample population under study for coronary heart disease. American Journal of Epidemiology, 83, 537547. Clark, N.M., & Rokowski, W. (1983). Family caregivers of older adults: Improving helping skills. The Gerontologist, 23, 637-642. Clipp, E.C., & George, L.K. (1990). Psychotropic
drug use among caregivers of patients with dementia. Journal of the American Geriatrics Society, 38, 227-235.
Cooreman, J., Thom, T.J., & Higgins, M. W. (1990). Mortality from chronic pulmonary diseases and asthma in France, 1969-1983. Comparisons with the United States and Canada. Chest, 97, 213-219. Croog, S . , & Fitzgerald, E. (1978). Subjective stress and serious illness on a spouse: Wives of heart pa-
262
RESEARCH IN NURSING & HEALTH
tients. Journal of Health and Social Behavior, 19, 166- 178. Derogatis, L.R. (1977). SCL-90-R Administration, scoring and procedures manual-I. Baltimore: Johns Hopkins University School of Medicine. Derogatis, L.R., Meyer, J.K., & Diny, K.M. (1981). Psychopathology in individuals with sexual dysfunction. American Journal of Psychiatry. 138, 757763. Derogatis, L.R., Rickels, K., & Rock, A.F. (1976). The SCL-90 and the MMPI: A step in the validation of a new self-report scale. British Journal of Psychiatry, 128, 280-289. Dillehay, R.C., & Sandys, M.R. (1990). Caregivers for Alzheimer’s patients: What we are learning from research. International Journal of Aging and Human Development, 30, 263-285. Enright, R.B. (1991). Time spent caregiving and help received by spouses and adult children of brainimpaired adults. The Gerontologist, 31, 375-383. Fortin, M-F., Coutu-Wakulczyk, G., & Engelsmann, F. (1989). Contribution to the validation of the SCL-90-R in French-speaking women. Health Care for Women International, 10, 27-41. Given, B., Stommel, M., Collins, C., King, S., & Given, C.W. (1990). Responses of elderly spouse caregivers. Research in Nursing & Health, 13, 7785. Gottlieb, B.H., & Selby, P.M. (1989). Social support and mental health: A review of the literature (Research Rep.). Guelph, ON: University of Guelph. Haley, W.E., Levine, E.G., Brown, S.L., & Bartolucci, A.A. (1987). Stress, appraisal, coping, and social support as predictors of adaptational outcome among dementia caregivers. Psychology and Aging, 2 , 323-330. Hirsch, B. (1979). Psychological dimensions of social networks: A multi-method analysis. American Journal of Community Psychology, 7, 263-277. Horowitz, A. (1982, November). Predictors of caregiving involvement among adult children of the frail elderly. Paper presented at the 34th Annual Meeting of the Gerontological Society of America, Boston, MA. House, J. (1981). Work stress and social support. Reading, MA: Addison-Wesley. Jutras, S . , & Veilleux, F. (1991). Informal caregiving: Correlates of perceived burden. Canadian Journal on AginglLa Revue canadienne du vieillissemeni, 10, 40-55.
Lapierre, L., & Adams, O.B. (1989). Leu personnes rigkes et la santk, Fa va? [Whither health and older people?] (Rapport de I’enqutte SantC QuCbec. tome I ) [Research Rep. Health Survey in Quebec, volume I]. QuCbec: Les Pub!ications du QuCbec. Lauzon, S. (1988). Etat de santk des conjointssoignants des malades atteints de dkrnence du type Alzheimer et facteurs associds [The health of spouse caregivers of persons with Alzheimer disease and associated factors]. Unpublished master’s thesis, Universitt de MontrCal, MontrCal, QuCbec.
Lazarus, R.S., & Folkman, S. (1984). Stress, appraisal, and coping. New York: Springer. Levasseur, M. (1983). Des problkmes prioritaires: La maladie selon les dges de la vie [Priority problems: Illness according to age]. QuCbec: Gouvernement du QuCbec. Uvesque, L., & Cossette, S. (1991). Revue critique d’Ctudes sur le soutien social et sa relation avec le bien-ttre des aidants de personnes atteintes de dCmence [Critical review of studies on the relationship between social support and well-being of caregivers of demented persons]. Revue canadienne de santk mentale communautairelCanadian Journal of Community Mental Health, 10, 65-93. Mace, N.L., & Rabins, P.V. (1981). The 36-hour day. Baltimore: The Johns Hopkins University Press. Manfreta, J., Mao, J., & Litven, W. (1989). Morbidity and mortality from chronic obstructive pulmonary disease. American Review of Respiratory Disease, 140, S19-26. McSweeny, J.A., Grant, I., Heaton, R.K., Adams, K.M., & Timms, R.M. (1982). Life quality of patients with chronic obstructive pulmonary disease. Archives of Internal Medicine, 142, 473-478. Montgomery, R.J.V., Gonyea, J.G., & Hooyman, N.R. (1985). Caregiving and the experience of subjective and objective burden. Family Relations, 34, 19-26. Murrell, S.A., Norris, F.H., & Hutchins, G.L. (1984). Distribution and desirability of life events in older adults: Population and policy implications. Journal of Community Psychology, 12, 30 1-3 1 1 . Norusis, M.J. (1990). SPSSIPC + 4.0 Base manual. Chicago: SPSS. Oberst, M.T., Thomas, S.E., Gass, K.A., & Ward, S.E. (1989). Caregiving demands and appraisal of stress among family caregivers. Cancer Nursing, 12, 209215. Phillips, L.R., Rempusheski, V.F., & Morrison, E. (1989). Developing and testing the beliefs about caregiving scale. Research in Nursing & Health, 12, 207-220, Prigatano, G.P., Wright. E.C., & Levin, D. (1984). Quality of life and its predictors in patients with mild hypoxemia and chronic obstructive pulmonary disease. Archives of Internal Medicine, 144, 16131621. Pruchno, R.A., & Resch, N.L. (1989). Husbands and wives as caregivers: Antecedents of depression and burden. The Gerontologist, 29, 159-165. Reeder, L.G., Schrama, P.G.M., & Dirken, J.M. (1973). Stress and cardiovascular health: An international cooperative study - l . Social Science and Medicine, 7 , 573-584. Schulz, R., Visintainer, P., & Williamson, G.M. (1990). Psychiatric and physical morbidity effects of caregiving. Journal of Gerontology, 45, P181-191. Sexton, D.L., & Munro, B.H. (1985). Impact of a husband’s chronic illness (COPD) on the spouse’s life. Research in Nursing & Health, 8 , 83-90. Sheehan, N.W., & Nuttall, P. (1988). Conflict, emo-
CAREGIVING TASKS / COSSETTE AND LEVESQUE
tion, and personal strain among family caregivers. Family Relations, 37, 92-98. Steinmetz, S.K., & Amsden, D.J. (1983). Dependent elders, family stress, and abuse. In T.H. Brubaker (Ed.), Family relationship in later life (pp. 173193). London: Sage. Stoller, E.P., & Pugliesi, K.L.(1989). Other roles of caregivers: Competing responsibilities of supportive resources. Journal of Gerontology, 44, S23 1-238. Stommel, M., Given, C. W., & Given, B. (1990). Depression as an overriding variable explaining caregiver burdens. Journal of Aging and Health, 2 , 81102.
263
& Sullivan, L. (1992). Tennstedt, S., Cafferata, G.L., Depression among caregivers of impaired elders. Journal of Aging and Health, 4 , 58-76. Zarit, S.H. (1989). Issues and directions in family: Intervention research. In E. Light & B.D. Lebowitz (Eds.), Alzheimer’s disease treatment and family stress: Directions for research (pp. 458-486). Rockville, MD: National Institute of Mental Health. Zarit, S.H., Orr, N.K., & Zarit, J.M. (1985). The hidden victims of Alzheimer’s disease: Families under stress. New York: University Press.
