| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
| ||||||||||||||||||||||||||||||||
RESEARCH ARTICLE |
a Mental Illness Research, Education and Clinical Center (MIRECC), Veterans Affairs Palo Alto Health Care System, California
b Stanford University School of Medicine, Palo Alto, California
c Family Studies Research Program, Veterans Affairs Palo Alto Health Care System, California
d Pacific Graduate School of Psychology, Palo Alto, California.
Dolores Gallagher-Thompson, Mental Illness, Education, & Clinical Center (MIRECC), Mail Code: 182B, Veterans Affairs Palo Alto Health Care System, 795 Willow Rd., Menlo Park, CA 94025 E-mail: dolorest{at}stanford.edu.
 |
Abstract |
|---|
 TOP Abstract MethodsResultsDiscussionReferences |
|---|
Methods. Twenty-seven wives caring for moderately impaired husbands with probable AD and 27 noncaregiving wives from comparable sociodemographic backgrounds were videotaped in their homes during both mealtime and a future event planning task. In addition, they completed self-report questionnaires to assess depression, stress, relationship mutuality, and perceived hope.
Results. Compared with their counterparts, caregiving wives reported higher levels of depression and stress, but similar shared values and closeness. For the three factors developed from the Marital Interaction Coding System (MICS, version IV; Supportive Facilitative, and Rapport Building), a complex pattern of results was found in which disease status, type of task, and gender interacted significantly. Noncaregiving couples were more interactive overall and expressed more support to each other. Caregiving wives were found to be most facilitative during the planning task, whereas AD husbands were highest on interactions that built rapport (e.g., smiling) during that same task. Caregiving wives actually increased their facilitative behavior from the mealtime to planning task, probably reflecting the increased demand characteristics of the latter.
Discussion. This study is one of a small body of literature to describe the negative impact of AD on spousal communication as observed and coded in two videotaped interaction situations in the home. Suggestions are made for future research, including the inclusion of longitudinal designs and non-Caucasian couples.
RESEARCH over the past two decades has shown that, after declines in marital closeness and satisfaction in midlife, marriages become increasingly strong and positive as couples enter old age (
Brubaker 1990
;
Carstensen, Gottman, and Levenson 1995). For example, in an observational study of emotional behavior in noncaregiving older couples in long-term marriages,
Carstensen and colleagues 1995 found that such couples expressed less negativity and more affection toward one another than middle-aged couples did when discussing highly charged marital problems. Other research has focused on key characteristics of successful long-term marriages.
Lauer and Lauer 1986 and
Lauer, Lauer, and Kerr 1990, among others, found that happy couples in long-term marriages approached their everyday problems with a problem-solving orientation that stressed attacking the issue rather than the mate, maintaining calmness and flexibility, and keeping issues in perspective. Both of these studies also found that emotional intimacy played a key role in determining happiness—95% of the 100 couples assessed in the study by
Lauer and colleagues 1990 study stated that their spouse was their best friend. Other studies have found that couples married more than 30 years generally report greater feelings of closeness (
Stinnett, Carter, and Montgomery 1972) and higher levels of emotional support (
Erikson, Erikson, and Kivnick 1986) than younger couples. Given these data, a question can be raised as to how marital satisfaction is likely to be affected when one spouse is significantly cognitively impaired, and therefore unlikely to be able to problem solve well or to provide emotional intimacy to his or her mate. As the absolute number of persons with dementing illnesses continues to rise, this question gains importance, particularly for women who are spousal caregivers.
Women provide most of the informal, hands-on, home-based care that delays or prevents institutional placement of demented persons (
Stone, Cafferata, and Sangl 1987); marital dissatisfaction on their part could lead to premature abdication of the caregiving role. Note that most prior research on marital satisfaction with "well" couples has relied on global self-report measures, which could be influenced by social desirability factors or cognitive dissonance. Direct observation of interaction patterns could provide valuable information about crucial links between interpersonal behaviors and reported satisfaction for older couples. Furthermore, directly comparing couples in which one spouse is demented with couples without a demented spouse may identify common interpersonal interactions that promote marital satisfaction. Following is a summary of current knowledge about the impact of dementia on caregivers and on couples' communication patterns and reported marital satisfaction.
Increased stress and burden are typically reported by family caregivers of dementia patients. Caregiving is often associated with reported declines in physical and mental health, as well as social functioning of the caregiver (
Gallagher, Rose, Rivera, Lovett, and Thompson 1989;
George and Gwyther 1986;
Gilhooly 1989;
Haley, Brown, and Levine 1987;
Pruchno and Resch 1989;
Schulz, O'Brien, Bookwala, and Fleissner 1995). Attempts to identify specific characteristics of caregiving detrimental to health and well-being have emphasized the quality and dynamics of the caregiver–care recipient relationship (
Niederehe and Scott 1987), as well as the affective ties between the caregiver and care-recipient (
Chenoweth and Spencer 1986;
Deimling and Bass 1986;
Fitting, Rabins, Lucas, and Eastham 1986;
Zarit, Todd, and Zarit 1986). On the other hand, some studies have found that characteristics of the caregiver–care recipient relationship do not influence the caregiver's feelings about his or her role (
Jenkins, Parham, and Jenkins 1985;
Zarit, Reever, and Bach-Peterson 1980), leaving some to suggest more careful investigation of interpersonal factors (
Birkel 1987;
DeLongis and O'Brien 1990).
Following a diagnosis of probable or possible Alzheimer's disease (AD), married couples enter a period of adjustment to the diagnosis (
Wilson 1989). As early as 6 months after the diagnosis, many spouses report that their relationship has deteriorated from one of close companionship to one of disorganization and loss of intimacy (
Blieszner and Shifflett 1990). Given the memory problems and aphasia characteristic of persons with AD, many spouses report that the biggest challenge following the diagnosis is communicating with their cognitively impaired spouse (
Chesla, Martinson, and Muwaswes 1994). They report feelings such as "we cannot sit down and talk now" and "he is not the same person he was." Common problems include the spouse with AD having difficulties finding the right word, understanding directions, sustaining conversations, or repeating ideas. Less common are increased talkativeness and failure to understand humor (
Bayles and Tomoeda 1991;
Ripich and Terrell 1988). In the early stages of the disease, these problems, while annoying, do not prevent effective communication from occurring. In fact,
Ripich, Carpenter, and Ziol 1997 and
Ripich, Vertes, Whitehouse, Fulton, and Ekelman 1991 found that people in the early and middle stages of AD upheld the requirements of orderly conversations; each person had a chance to talk, only one person talked at a time, and gaps between turns were brief. While this format was upheld,
Ripich and colleagues 1991 observed that persons with AD take more turns talking, but use fewer words each turn. They theorized that brief turns placed less demand on the AD person's memory and allowed him or her to remain engaged in the conversation.
In addition to impairing basic communication skills, AD affects the person's capacity for self-consciousness, organized thought, and shared meaning that are the essence of human interaction and communication. The loss of these abilities has been shown to produce tension and resentment in the marital relationship (
Barusch and Spaid 1996;
Gladstone 1995;
Wright 1991). Barusch and Spaid found that dementia caregivers described their marriages as disengaged and reported considerably less closeness to their spouse than before the onset of dementia.
Wright 1991 compared 30 AD couples with 17 non-AD control couples and found that control couples reported handling tension-producing situations by openly expressing frustration while showing respect for the other's feelings, whereas AD couples were unable to do so. Because of the patient's deteriorated communication skills, differences could not easily be resolved through discussion. However, continued efforts to initiate and maintain positive interactions with the spouse can be beneficial to both caregiver and care recipient alike (
Wright 1994). Despite the potential importance of positive dyadic interactions for improved quality of life, caregivers seldom receive guidance in how to maintain maximal functional communication with their impaired spouse. In fact, spouses may unknowingly thwart communication and hasten the patient's linguistic decline. For example,
Ramanathan-Abbott 1994 demonstrated that open-ended questions, which encourage patients to recall past events and organize their narrative, along with a slowed pace of conversation and hesitation to complete sentences or to find words for patients, may help persons with AD retain communicative ability and confidence.
For the most part, descriptions of the caregiver–care recipient relationship have been based on caregiver self-report data, with only a few attempts to employ direct observation methods to provide corroboration.
Niederehe 1990 evaluated interaction patterns of caregiver-patient dyads in the laboratory setting and reported that caregivers who were more depressed displayed fewer positive and more negative interactions than those who were less depressed (
Niederehe 1990).
Shields 1992 successfully videotaped interactions of caregivers and other family members in the home and reported that angry and sad interactions by other family members toward the caregiver accounted for 44% of the variance in caregiver depression. However, the actual persons with AD were not included in any of the interactions observed in that study.
Gallagher-Thompson and associates 1997 have demonstrated the feasibility of videotaping caregiver-spouse and care recipient dyads in their homes while engaged in both nonstructured mealtime and structured planning activities. Preliminary data with 15 couples suggested a positive relationship between distress and marital conflict, as well as an inverse relationship between distress and the frequency of positive interactions. However, in that study, no data were obtained from noncaregiving marital dyads for comparison. In the present study, besides adding this comparison group, some refinements were made in how instructions were given for each task, based on feedback from the participants. Otherwise, no substantive changes were made in the procedures.
One aim of the present study was to explore the relationship between interaction patterns and several variables reflecting positive thoughts and perceptions, including caregiver's levels of resilience and hope (
Aspinwall and Taylor 1997) and the perceived mutuality of the marital relationship (
Archbold, Stewart, Greenlick, and Harvath 1990). Resilience refers to a person's ability to respond flexibly and appropriately to stressful situations; one common way to assess this construct is to measure hope. Mutuality refers to several aspects of marital satisfaction, including reciprocity in the relationship, feelings of closeness, and shared values and everyday pleasures. These variables were selected because prior research has suggested one or more of these may serve a protective function and actually mitigate typical caregiver role strain and stress (
Archbold et al. 1990). Our main interest, however, was to directly observe marital interactions of two groups of couples: one in which the wife is a caregiver for her husband and one in which the wife is not a caregiver. Videotaped observations captured interactions between couples in which the husband had been diagnosed with AD and noncaregiving control couples of comparable age and socioeconomic status. Because communication patterns may vary as a function of situational demands, couples were videotaped in two distinctly different but customary situations: during mealtime and while planning a future event. Because we wanted two relatively distinct and nonthreatening situations, we decided against other tasks, such as putting block designs together (
Cavanaugh et al. 1989), which were known to be more stress inducing.
This study was designed to address three primary questions: (a) Would there be differences in the interaction patterns between caregiver–care recipient couples and comparison control couples? (b) Would any observed differences vary as a function of task demands? And, (c) would the responses of caregivers in these tasks be correlated with levels of depression and overall stress? A secondary aim was to explore relationships between interaction patterns and indices of hope and resilience, along with mutuality as an index of satisfaction in the marital relationship, in order to suggest possible future research directions.
|
Methods |
|---|
|
TOPAbstractMethodsResultsDiscussionReferences |
|---|
AD couples were recruited from the Clinical Core of the ADDTC where, following diagnosis, they were studied longitudinally in a variety of research projects (e.g., sleep disturbances, patient depression, responsiveness to treatment). All patients enrolled in this center had a diagnosis of "probable Alzheimer's disease" based on NINCDS-ADRDA criteria (
McKhann et al. 1984). In addition, patients with Mini-Mental Status Examination (MMSE;
Folstein, Folstein, and McHugh 1975) scores between 15 and 25 at the time of entry into the ADDTC, who were functionally impaired (as reported by their family member) in at least two common activities of daily living (e.g. driving, or handling finances), and who were being cared for by a family member who lived with them were invited to participate in this video project. Of those eligible (N = 84), 45% agreed to the videotaping (n = 38). Those who refused most often cited discomfort with the idea of being videotaped in their home or busy schedules that did not permit the time commitment.
For this article, we chose to focus only on wife caregivers (n = 27) because prior research has found differences in how caregiving is experienced and how stressful it is perceived to be between husbands and wives (
Fitting et al. 1986;
Lieberman and Fisher 1995). Thus, not all dyads who were videotaped are included in this report. Noncaregiving couples were recruited through mailings, flyers, and newspaper advertisements. Respondents were screened by telephone. Couples who met the following criteria were included in the study: (a) the couple was married and living together; (b) both were aged 50 years or older; (c) neither was now or had been a caregiver to the other due to chronic illness; and (d) neither had been a caregiver to another individual within the past year.
The two samples were quite comparable in terms of demographic characteristics. Caregiving wives had a mean age of 69 and were caring for husbands with a mean age of 74 years. Noncaregiving wives had a mean age of 70 and their husbands had a mean age of 73 years. The mean length of marriage for AD couples was 41 years, with a range of 5–61 years (a very few, n = 3, being second or late-life marriages). Non-AD couples had a mean of 37 years of marriage, with a similarly wide range of 5–66 years. Caregiving and noncaregiving wives were well educated, with a mean of 14.6 years of schooling for caregiving wives and 15.6 years for noncaregivers. When evaluated statistically using t tests, no significant differences were found between groups in these variables. In terms of employment and income factors, more than a third of the wives in each group had served in professional or managerial roles when they were working. Slightly more noncaregivers than caregivers classified themselves as housewives (12 vs. 8) and almost half of the couples reported an average household income of more than $50,000. When these measures were evaluated using chi–square tests, no significant differences were found between the two groups. Taken together, these data indicate that this was a middle- to upper middle–class, Caucasian, volunteer sample. AD husbands were moderately impaired at the time of the videotaping, with an overall mean MMSE score of 20.3 and significant functional impairment; overall, they were viewed as being in the middle stage of the illness.
Measures
All participants were interviewed and videotaped by trained, master's-level research assistants. The interview consisted of obtaining informed consent to participate in the project, followed by demographic questions and a series of exploratory questions on social support, which are not reported here. In addition, all husband-and-wife dyads participated in a home videotaping session scheduled within two weeks of the initial interview.
Wives also completed a self-report packet at home and returned it by mail within a week of the interview. This was done to reduce response bias that can occur when answering direct questions about one's thoughts, feelings, and hopes, and to provide some privacy for responses. The packet contained several standardized self-report measures; for purposes of this study, analysis will be limited to the following four measures (alpha levels in this sample are given in parentheses):
), a 30-item yes/no self-rated measure to assess severity of symptoms of depression in older adults (Cronbach's 
= .94). Higher scores indicate more depressive symptoms. ), a 14-item, 5-point measure of self-perceptions of generalized stress that is not specific to caregiving (Cronbach's = .88). Higher scores indicate greater stress. ,
Aspinwall and Taylor 1997), a 12 item, 4-point measure that assesses the respondent's degree of agency (goal-directed determination) and pathways (planning of ways to meet goals). This measure has been used in prior research as an index of the construct of resiliency (Cronbach's = .80). Higher scores indicate more resiliency/hope. ), a 15-item measure that assesses the relationship between married couples on four subscales: shared values ( = .73), affective closeness ( = .90), shared pleasurable activities ( = .89), and reciprocity ( = .89). Higher scores indicate a greater degree of the construct being measured. In order to reduce the number of variables to be used subsequently as dependent measures, a principal components analysis with varimax rotation was performed on this set of psychosocial measures. Two factors were obtained, accounting for 75% of the variance. The first factor (44%) was comprised of the four mutuality scales, and the second (31%) included the GDS and PSS, which loaded negatively, and the HS, which loaded positively. All variables had weights over .750 on their respective factors. The first factor was labeled as Closeness and the second as Distress. Scores on these two factors were then used to compare caregiving and noncaregiving wives.
Videotaping Procedure
Videotaping took place at the dining table in the couple's home. A compact video camera (with remote microphone) was placed, inconspicuously, several feet from the dining area, and was turned on before the couple sat down to their meal. Participants were first videotaped during an afternoon or evening meal with no particular emphasis placed on any specific communication target except events of the day or other issues that might arise during the course of the meal. A brief period of testing the equipment and encouraging participants to relax and speak and behave naturally followed. Specific instructions for the mealtime task were then given: "I am going to leave the room now. Please enjoy your meal as you normally would. Do not hesitate to get up from the table briefly if you need something. I'll come back in 15 to 20 minutes."
After the meal was over, the researcher briefly asked about the taping experience, making sure the couple felt comfortable proceeding to the more-structured planning task, in which the couple worked together to plan a brief "outing" to visit friends or engage in some other activity that would be mutually pleasant. The tasks were chosen on the advice of one of the authors (T. J.) who had used them in prior research with dysfunctional couples. Instructions for the second task were as follows: "Now I'm going to ask you to plan something for the two of you to do sometime in the near future. It could be a visit to a friend, a shopping trip, or a longer getaway. I would like you two to talk about it and discuss any relevant issues, including what you would like to do, when you would like to do it, and what you might need to bring, including any details that you feel need to be taken care of. I will come back in 10 to 15 minutes."
At the conclusion of the planning segment, the video camera was turned off and couples were debriefed and encouraged to discuss feelings or questions they had about the taping. For the most part, couples commented that their videotaped interactions had been typical, and that following brief discomfort with the camera's presence, they were able to talk and relate as they normally did.
Video data were coded by staff of the Oregon Marital Studies program at the University of Oregon, using the Marital Interaction Coding System (MICS-IV;
Weiss and Summers 1983). The MICS-IV was selected because it is the most frequently used marital interaction coding system for psychological research, and it has been widely applied to families with other types of significant chronic problems, including alcoholism (
Jacob and Leonard 1992) and schizophrenia (
Leff and Vaughn 1985). Given that no unique coding system presently exists for scoring interactions of AD patients and their families, we believed that the MICS-IV was a reasonable choice. Inter-rater reliability was computed by the Oregon raters on 20% of our tapes and ranged from .70 to .83. The MICS-IV system provides frequency counts for the following 37 categories of verbal and nonverbal communication: agree, approve, accept responsibility, assent, attend, command, comply, criticize, compromise, show dysphoric affect, disengage, disagree, disapprove, deny responsibility, excuse other, excuse, humor, interrupt, mindread, mindread positive, noncompliance, offer negative solution, not tracking, off topic, problem external, problem internal, positive physical contact, paraphrase, positive solution, put down, question, smile, laugh, talk, turn off, threaten, and hostile voice tone.
MICS-IV Data Reduction
Preliminary inspection of descriptive statistics incorporating all 37 codes indicated that the gross MICS-IV profile for the two groups was similar in both the kind and the number of interactions. More detailed scrutiny revealed that 9 scoring codes were not used for any of the groups (disapprove, deny responsibility, excuse other, excuse, noncompliance, offer negative solution, threaten, hostile voice tone, and show dysphoric affect), and 11 of the codes were used only once or twice by a few subjects (approve, accept responsibility, command, comply, criticize, compromise, disengage, mindread, positive physical contact, put down, and turn off). There was no preponderance of these codes within any of the groups. Most of these no- and low-occurrence codes were negative in character (e.g., disapprove, criticize). Thus, at first glance, one could characterize the interaction pattern of the four groups as being similar in terms of the types and structure of the communication patterns, although the substance and quality of the communications are noticeably different.
This observation prompted us to review the distributional characteristics of the remaining codes carefully before proceeding with further analyses. Distributions were examined for the 17 codes remaining and transformations were made based on each code's unique set of characteristics as recommended by
Tabachnick and Fidell 1989. Z scores were then calculated in order to make intercode comparisons, and a principal components analysis with varimax rotation was performed to further reduce the data. Three factors, accounting for 62.3% of the total variance, were extracted for use in further analyses. Factor 1 contains assenting and paraphrasing behaviors such as expressing agreement or expressing approval. We called it the Supportive factor. Factor 2 contains such codes as "propose a positive solution," "make a clarifying statement about the matter currently being discussed," and "infer positive intentions in the spouse," and was labeled by us as the Facilitative factor. Items in Factor 2 reflect sophisticated behaviors, requiring multidimensional thinking and organized thought. Factor 3 consists of such codes as "smiling," "laughing," "using humor," and "attending closely to the other person's speech." This factor we called Rapport. (Although a prior factor analysis was reported by
Heyman, Eddy, Weiss, and Vivian 1995, we were not able to replicate those results in the current sample and therefore their results were not used).
Data Analysis
Differences between caregiver and noncaregiver wives on the closeness and distress psychosocial factors were evaluated using two-sample, independent t tests. For each of the three MICS-IV factors (i.e., Supportive, Facilitative, Rapport) differences between caregiver and control couples were evaluated, using multivariate analysis of variance for repeated measures (MANOVA) with 3 variables: disease status (2 levels), gender (2 levels) and task (2 levels). Finally, correlations between select psychosocial measures and the three MICS-IV factors were also completed in order to evaluate their degree of association.
|
Results |
|---|
|
TOPAbstractMethodsResultsDiscussionReferences |
|---|
). Noncaregiving wives had a significantly lower depression mean of 2.4 (t at 1/54 df = 5.14, p < .0001). Caregivers' perceived stress scores were moderately high, with a mean of 22.3 out of 56, while control wives reported significantly less stress, with a mean of 15.3 (t at 1/54 df = 3.44, p < .001). Caregivers' overall stress level in this study was comparable to values obtained in other caregiving research conducted by our group (
Lovett and Gallagher 1988).
|
for noncaregivers. Finally, on subscales of the Mutuality Scale, there was no difference on self-reported closeness to their husbands (p = .315) or shared values (p = .138); however, caregiving wives reporting significantly less reciprocity in their marriages (p < .001) and fewer shared pleasurable activities with their husbands noncaregiving wives ( p < .001). These results are comparable to the findings of
Archbold and colleagues 1990.
Videotaped Interaction Behaviors
In general, the planning task was more interactive than the mealtime exchange, producing more total codes than the mealtime task. Of the planning codes, a greater proportion were from the facilitative factor than were seen in the mealtime. Because of the legitimate and revealing differences between the tasks, they were kept separate for our analyses.
Table 2 provides the means and standard deviations for the three factors obtained from the MICS for each of the two tasks, grouped by caregiving and relationship status. In order to address the first two research questions (Would there be differences in interaction patterns between caregiving and noncaregiving couples? Would any observed differences vary as a function of task demand?), these data were subjected to a 2 (disease status) x 2 (gender) x 2 (type of interaction task) multivariate analysis of variance (MANOVA) for repeated measures, with task as the repeated measure (see Table 3 ). Four couples had to be dropped from this analysis because of technical problems with the videotaping of either segment. Estimates of Wilks's Lambda in the MANOVA indicated significant main effects for disease, gender, and task. The univariate analyses indicated that these effects were most prominent in Factor 2, facilitative interactions, and to a lesser extent in Supportive interactions (Factor 1). There were no differences in these main effects or their interactions on Factor 3, reflecting rapport-building interactions.
|
|
In terms of gender effects, women were in general more facilitative than men (irrespective of disease status or task), but there was no gender effect on either the supportive or rapport-building factors. Facilitative and supportive interactions were higher in the planning task than in the mealtime. None of the two-way interactions were significant, but the three-way interaction (Task x Disease x Gender) was highly significant for Factor 2 and also significant for Factor 1, but not significant for Factor 3 (see Table 3 ).
Inspection of the univariate analyses indicates that the strongest effect is on the Facilitative factor (Factor 2). Fig. 1 provides a graphic illustration of the three-way interaction for this factor. The caregiving wives and control husbands had a greater increase in facilitative responses than their counterparts from the mealtime task to the planning task, t(23) = 5.785, p < .000, and t(25) = 6.235, p < .000, respectively. However, the change for husbands with dementia only approached statistical significance, t(23) = 2.033, p = .054. Further post-hoc testing revealed no difference between the two husband groups during the mealtime task, t(48) = 2.101, p = .082 with a Bonferroni correction applied, whereas the difference between these two groups was substantial for the planning task, t(48) = 5.772, p < .000. Conversely, the two wife groups were substantially different during the mealtime task, t(48) = 3.463, p = .002, but their factor scores were nearly identical during the planning task, t(48) = .656, p = 1.000.
|
No significant differences were found for Factor 3, reflecting rapport-building interactions. However, control wives scored slightly higher in laughter and use of humor during the mealtime than caregiving wives. During the planning session, both groups of wives displayed less laughter and humor than they did during the meal, and caregiver wives showed the least rapport of all four groups (see Table 2 means).
In order to address the third research question of interest (namely, the degree of association between caregivers' psychosocial variables and MICS-IV interaction factors), Pearson correlation coefficients were computed for the caregiver wives only and are displayed in Table 4 .
|
|
Discussion |
|---|
|
TOPAbstractMethodsResultsDiscussionReferences |
|---|
and
Shields 1992, who were among the first to use the direct observational technique of videotaping to access verbal and nonverbal interactions between AD patients and their caregivers (Niederehe) or between persons with AD and their family members (Shields). In the current study we built upon their methodology by including the following methods in our research design: (a) videotaping was done in subjects' homes to approximate real life more closely; (b) we controlled for gender and relationship differences by selecting only wife caregivers of husbands diagnosed with probable AD; (c) we also recruited a sociodemographically comparable sample of wives who were not caregivers to their spouses to assess more clearly disease-related differences in interaction style; (d) two different types of everyday situations were videotaped (mealtime and future event planning) in order to assess how interaction patterns were similar or different across tasks; (e) a well-validated system for coding marital interaction was used—the MICS-IV—which, although not previously used with this population, is the system most often employed in research of this type; and, (f) some additional psychosocial measures (of resiliency and perceived mutuality in the marriage) were added to explore their relationship with interaction style. Following data reduction of both the psychosocial measures and the MICS-IV codes, a variety of analyses were conducted to address the three questions of interest that guided this research: Would there be differences in interaction patterns between caregiving and noncaregiving couples? Would any observed differences vary as a function of task demands? And would caregivers' interaction patterns correlate with other variables likely to affect results, including depressive symptoms, overall stress, resiliency, and perceived mutuality of the marital relationship?
Regarding the first two questions, results were complex, as evidenced by a significant three-way interaction among disease status, gender, and task on one of the three factors obtained from principal components analysis of usable MICS-IV codes—namely, the Facilitative factor (Factor 2). This has to be interpreted in context, along with other significant results on the Supportive factor (Factor 1) and lack of any significant findings on the Rapport-Building factor (Factor 3). The findings point to both differences and similarities between the sample's four subgroups; taken together, they highlight the impact of AD on these couples' patterns of communication. First, noncaregiving couples were significantly more interactive during both tasks. This is understandable, given that most people with AD in the middle stage of the disease have difficulty verbalizing and expressing themselves. In accord with descriptions of AD communication provided by
Ripich and colleagues 1991 and
Ripich and Terrell 1988, fewer ideas were exchanged between our AD husbands and their wives, and those ideas were expressed in a briefer and more simplified manner by both the verbally challenged husband and the caregiving wife.
Second, analyses examining the wives alone found that caregiving wives were less supportive of their husbands' ideas than noncaregiving wives. Instead, caregiving wives invested their communication energy into solving problems and clarifying topics of conversation for their husbands, as evidenced in their more frequent use of codes from the facilitative factor (Factor 2), which was particularly noticeable in the planning task.
Third, as anticipated, AD husbands utilized fewer codes from the supportive and facilitative factors (Factors 1 and 2, respectively) than the comparison group of husbands. This is not surprising given the recognized deterioration in communicative skills brought on by AD. Our AD husbands behaved very much in accordance with
Ripich and Terrell 1988 earlier descriptions of the midstage AD patient. They did uphold the requirements of orderly ordinary conversation in that they took turns speaking; in fact, they had no higher incidence of interrupting, talking off topic, or withdrawing from the conversation than nondemented husbands (it should be noted that these particular codes were not frequently used by either group). However, AD husbands used fewer words and seemed to have lost much of their ability to plan and propose long-range solutions during the planning task, even though they were asked to plan something relatively simple (a brief future outing of some kind). Yet, as found earlier by
Bayles and Tomoeda 1991 and
Ripich and Terrell 1988, AD husbands also showed good ability to use and understand humor. In fact their scores on the rapport-building factor (Factor 3) were highest of all four subgroups on the planning task, suggesting that AD husbands used humor to diffuse some of their discomfort with the challenges associated with this task for them. By including both AD and non-AD couples and by using two tasks, we believe we were able to detect more subtle aspects of communication than would have been observable without the comparison group.
Taken as a whole, these results suggest that, despite the progressive decline of AD, certain aspects of communication between long-term marital partners remain remarkably intact, particularly at mealtime, which seemed to be a pleasant experience for most of the couples (despite low frequency and simplicity of conversation). In the absence of longitudinal data, we cannot determine the extent to which these results reflect changes over time in marital interaction patterns due to AD. These cross-sectional data do, however, suggest several future research questions (which will be discussed).
Regarding the third question of interest (namely degrees of association between the particular psychosocial measures used and the observed interaction patterns), let us start by discussing the findings on the psychosocial variables themselves. We found significantly higher levels of depressive symptoms and generalized stress in the caregiving versus control wives, in agreement with much prior research (e.g.,
Baumgarten et al. 1992;
Dura, Stukenberg, and Kiecolt-Glaser 1991). We also examined caregivers' perceived levels of resilience (as indexed by the Hope Scale) and several aspects of mutuality in the spousal relationship (to assess marital satisfaction indirectly), neither of which had received much attention in prior research. We found, as expected, lower resilience among caregiving wives; this may be associated with greater difficulties coping as the demands of caregiving increase and the rewards associated with the situation begin to diminish (
Schulz et al. 1995). We also found less reciprocity and fewer shared pleasures (two dimensions of mutuality that were assessed by the particular measure we used). Interestingly, reports of both shared values and a sense of emotional closeness (assessed on the same measure) were equivalent between the two groups of wives, suggesting that these aspects of marital satisfaction were still present, despite the impact of AD on other aspects of their marital relationship. Perhaps these wives' ability to maintain their perception of some degree of emotional intimacy is a critical factor in enabling them to continue to maintain themselves in this role.
Overall, our results here confirm prior research—namely, that AD caregiver wives are more psychologically distressed than their counterparts of comparable age, socioeconomic status, and time married; this was true despite their ability to maintain a sense of emotional closeness with their partner and thus still perceive some sense of marital satisfaction. However, when these psychosocial measures were correlated with the factor scores reflecting observed interaction patterns, no significant relationships were found between depression or stress and any of the three interaction factors. This is contrary to the earlier work of
Niederehe 1990 and
Shields 1992 in which caregivers' depression significantly influenced their interaction style. Possibly our results are explainable by the low number of caregivers who were significantly depressed: Fewer than 20% scored at 20 or above on the GDS, which is the recommended cut-off score for moderate depression (
Yesavage et al. 1983). Of the exploratory variables that were included, only resiliency (as indexed by the Hope Scale) was correlated with any of the interaction factors such that higher caregiver resiliency was associated with more facilitative interactions during mealtime. As suggested by
Aspinwall and Taylor 1997, resilience may be a personal characteristic associated with more adaptive coping to ongoing stress. Future research to further explore its role as a possible moderator of caregiver function is suggested.
In addition to this study's strengths, several of its limitations should be noted. First, our sample was relatively homogeneous, consisting of well-educated Caucasian couples of middle to upper-middle socioeconomic status, not in great psychological distress, who volunteered to participate in this project. Those who were videotaped passed through several self-selection processes, including presenting themselves at the AD research center in the first place, agreeing to a personal interview, and permitting a researcher to enter their home with a video camera. This group represents approximately 45% of all couples participating in the center's larger research program. Second, we did not include a more typical measure of marital satisfaction as studies focusing on correlates of satisfaction in long-term marriages have done. We also did not use a longitudinal design, so we cannot be certain whether change in patterns of interaction had occurred. However, by including the noncaregiving control couples for comparison, we were able to detect current differences in both interaction patterns and aspects of martial satisfaction. Third, we did not have a comparison group of women caring for a spouse with a serious physical (but nondementing) illness. Therefore, we cannot be certain that the differences observed were not due to caregiving in general, rather than being unique to dementia caregiving.
Fourth, this study did not address ethnic or cultural differences in caregivers' experience, although such differences are known to exist and most likely have an impact on the nature of marital interactions (
Aranda and Knight 1997;
Connell and Gibson 1997). For example, a recent study by
Ripich, Kercher, Wykle, Sloan, and Ziol 1998 found that African American caregivers benefited more than White caregivers from training in how to improve their communication with a relative with AD. African Americans reported more positive affect and fewer daily hassles after training compared to their White counterparts. Further research is clearly needed on how families with non-Caucasian cultural and ethnic backgrounds interact with their care receivers.
Finally, the MICS coding system, while widely used in non–aging focused marital interaction research (e.g., with schizophrenics and alcoholics; see
Jacob and Leonard 1992) may not be the most appropriate for assessing quality of interaction in AD couples. Earlier we reported that both AD and control participants failed to use many of the MICS codes; AD couples had more no-occurrence codes than controls. For example, AD couples utilized just 5 of 15 possible negative codes, and controls used 7 of the 15. Among those not used were the most clearly negative ones such as putting down or directly criticizing the other person. It is possible that use of different tasks would have been more effective in eliciting stronger emotional responses from participants. For example,
Cavanaugh and colleagues 1989 successfully documented differences between AD couples and matched controls while each couple worked together to complete the more demanding task of assembling blocks into progressively more difficult patterns, where nonverbal reasoning capacity is needed to assemble the designs correctly. Different instructional strategies were used by the dementia caregivers including concrete, explicit directions and considerable praise to encourage continued responding. A unique coding schema was developed for that study that did contain one negative code labeled restriction (verbal interruption or saying "no") but that was used less than 10% of the time by the caregivers. It may be that outright negative interactions are few among couples who volunteer for these kinds of research studies. Only by using less intrusive methods to observe daily, routine interactions will we be able to learn more about the true frequency of negative interactions among caregiving couples.
Additional research is also needed on interaction patterns in other AD family members. Would adult daughters, for example, demonstrate more negative interpersonal interactions than wives due to their generally higher levels of life stress as members of the "sandwich generation" (
Brody 1981;
Dura et al. 1991)?
Lieberman and Fisher 1995 are among the first to include multiple generations in their self-report study of the impact of caregiving on members of 97 caregiving families. However, no observations of interactions were made in that study and, to our knowledge, no studies have yet been undertaken in which interactions were observed or videotaped by "position" in the family. Such research could have important implications for patient care and family stress (
Brubaker 1990).
Finally, future studies should include longitudinal designs to assess change over time in patterns of communication among older couples, with and without a spouse with dementia. Such information could inform efforts to improve older couples' quality of life, as was demonstrated recently by
Corbeil, Quayhagen, and Quayhagen 1999. They found that satisfaction with interactions with the demented relative changed over a period of 9 months for those who participated in their intervention program. Greater improvement over time was associated with caregivers' increased ability to use problem-focused coping strategies in the caregiving situation.
In sum, although observational research is difficult and labor intensive to conduct, it can lead to refinement of developmental theory about how long-term marriages weather chronic stress. This, in turn, should help us design appropriate interventions to reduce stress and improve caregiver and patient well-being. For example, caregiving wives could be expected to benefit from training and practice in how exactly to become more supportive in their interactions with their spouse during planning-type tasks. Several communication skills training approaches exist and have been used for this purpose in prior research (see, e.g.,
Bourgeois 1993,
Freeman, Barocas, and Chrosniak 1999;
Ripich 1994). Given generally high levels of depression and stress, this type of training might lead to more of a sense of closeness and of working together to achieve common goals, which can enhance wives' sense of emotional intimacy in their marital relationship. On the other hand, wives who are observed to be more critical of their spouses and negative toward them might benefit from a psycheducational approach in which they learned a better set of skills for coping with the frustrations of dementia caregiving (
Gallagher, Wrabetz, Lovett, DelMaestro, and Rose 1989;
Gallagher-Thompson and DeVries 1994). Additional research is needed to evaluate the impact of such programs on couples.
|
Acknowledgments |
|---|
Received for publication June 8, 2000. Accepted for publication November 7, 2000.
|
References |
|---|
|
TOPAbstractMethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  L. Garand, M. A. Dew, B. Urda, J. H. Lingler, S. T. DeKosky, and C. F. Reynolds Marital Quality in the Context of Mild Cognitive Impairment West J Nurs Res, December 1, 2007; 29(8): 976 - 992. [Abstract] [PDF]
|
|
|
|
 |
|
 J. A. Small, J. Perry, and J. Lewis Perceptions of family caregivers' psychosocial behavior when communicating with spouses who have Alzheimer's disease American Journal of Alzheimer's Disease and Other Dementias, September 1, 2005; 20(5): 281 - 289. [Abstract] [PDF]
|
|
| ||||||||||||||||||||||||||||||||
| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
