ABSTRACT CARING FOR A PERSON WITH DEMENTIA: EXPLORING TIME USE WITH TIME DIARY METHODOLOGY. by Sarah J. Hahn

Transcription

1 ABSTRACT CARING FOR A PERSON WITH DEMENTIA: EXPLORING TIME USE WITH TIME DIARY METHODOLOGY by Sarah J. Hahn Time diaries were collected from 33 caregivers caring for a relative with dementia. The diaries provided a retrospective look at the caregivers past 24 hours, detailing the activities performed, with whom caregivers spend their time, and their overall affect. Diaries were analyzed to document the caregivers overall time use while with the person with dementia, as well as the relationship between time use and key components of the transactional process model of stress (i.e., background and contextual characteristics, Hassles, Loss of Self, Role Competence, Capability, Role Captivity, and caregiving well-being). Results suggested that caregivers spent less time caring for their relative than they perceived. In addition, background and contextual characteristics (i.e., gender, kinship, occupational status, relative s level of cognitive impairment) as well as caregiver Capability, Loss of Self, and depressive symptomatology were associated with time use. The results of this study provide important implications for future research, policy, and practice.

2 CARING FOR A PERSON WITH DEMENTIA: EXPLORING TIME USE WITH TIME DIARY METHODOLOGY A Thesis Submitted to the Faculty of Miami University in partial fulfillment of the requirements for the degree of Master of Gerontological Studies Department of Sociology and Gerontology by Sarah Jane Hahn Miami University Oxford, Ohio 2012 Advisor: Dr. Jennifer Kinney Reader: Dr. Ronald Bulanda Reader: Dr. Heidi Ewen

3 Table of Contents Introduction Overview of Dementia and Caregiving 1 Overview of the Caregiving Stress Process 1 Time Use 2 Time Use among Caregivers to a Relative with Dementia 5 Method 8 Participants and Recruitment 8 Instruments Used in Data Collection 10 Results 15 Section 1: General Time Use 16 Section 2: General Time Use with the Person with Dementia 19 Section 3: Associations with Time Use - Background and Contextual Characteristics, Stressors, Loss of Self, Role Competence, Capability, Role Captivity and Well-being 23 Discussion 28 Appendix 1 37 Appendix 2 39 Appendix 3 40 Appendix 4 41 References 42 ii

4 List of Tables Table 1: Description of Caregivers and Persons with Dementia 9 Table 2: Caregivers Time Use: With Whom They Spend Time 17 Table 3: Description of Caregivers Time Use: Activity and Affect 18 Table 4: Description of Caregivers Time Use: With Person with Dementia 20 Table 5: Description of Caregivers Time Use: With Person with Dementia, 22 Activity by Affect Table 6: Impact of Background and Contextual Characteristics on Time Spent 23 with Person with Dementia Table 7: Length of Caregiving and DRS Scores on Time Use and Activity 25 Table 8: Impact of Caregiver Hassles, Loss of Self, Role Competence, Capability, 27 and Role Captivity on Overall Time Use and Activity iii

5 List of Figures Figure 1: Modified Transactional Stress Process Model 7 iv

6 Introduction Overview of Dementia and Caregiving As of 2008, 2.4 million women and 1 million men over the age of 71 had some form of dementia (Plassman, Langa, Fisher, Heeringa, Weir, & Ofstedal, 2007). Of the various types of dementia, Alzheimer s disease is the most prevalent form, accounting for 60 to 80 percent of cases. In 2012 alone, an estimated 5.4 million Americans suffer from Alzheimer s disease (Alzheimer s Association, 2012). Alzheimer s disease is a brain disease that causes problems with memory, cognition, and behavior. It interferes with daily activities, and worsens over time (Alzheimer s Association, 2012). The majority (87%) of individuals with dementia are cared for at home by a family or community member (Alzheimer s Association, 2009). This translates into more than 15 million Americans who provide unpaid care for a person with Alzheimer s disease or other dementia. The impact of caregiving is immense; caregivers risk having chronic health issues, poor emotional well-being, and physiological changes (Alzheimer s Association, 2012). Overview of the Caregiving Stress Process The literature on caregiving is vast, especially when it comes to caring for a person with dementia (e.g., Cambell, Wright, Oyebod, Crome, Bentham, Job, & Lendon, 2008; Pinquart & Sorensen, 2003; Mace & Rabins, 1999). Much of the research on informal caregiving for a relative with dementia is framed by the transactional process models of stress (e.g., Lazarus & Folkman, 1984; Pearlin, Mullan, Semple, Skaff, 1990). These models conceptualize caregiving as stressful (e.g., Pinquart & Sorensen, 2003) and frame the context in which the stress process unfolds and how the components of the process are interrelated. Lazarus and Folkman (1984) viewed the stress process as an antecedent-consequence approach. This model assumes that stress is not one factor, rather an organizing concept for understanding a wide range of phenomena (Lazarus & Folkman, 1984, pg. 11). They define stress as a relationship between the individual and her/his environment that is appraised by that individual as exceeding his or her resources, and therefore threatening her/his well-being. From this stress, the antecedent-consequence model identifies personality variables that can mediate the stressful event. The interaction of personality and environment, specifically how people appraise and cope with their stress and resultant outcomes, is the process of this transactional 1

7 perspective. This process is iterative; individuals are constantly re-appraising their environment, behavior, and resources to evaluate possible outcomes. Pearlin et al. (1990) approached the stress process as a culmination of different experiences and circumstances throughout time that vary from caregiver to caregiver. The multiple components of these experiences and circumstances are organized in four domains: the background and context of stress, stressors, mediators of stress, and the outcomes. Stress is a consequence of the interactions of these four domains and the components within each. Dozens of researchers have built programs of research using Lazarus and Folkman s (1984) and Pearlin et al. s (1990) conceptualizations to frame their investigations of the caregiving process (e.g., Mausbach, Roepke, Chattillion, Harmell, Moore, Romero-Moreno, & Grant, 2012; Kneebone & Martin, 2003;Semple, 1992). For example, using both stress process models as a guideline, Koerner, Shirai, and Kenyon (2010) found that as caregivers care-related stressors increase (e.g., performing ADLs, IADLs, managing problematic behavior on the part of the care-receiver), caregivers experience more depressive symptoms, feelings of burden, and physical health issues. Utilizing Pearlin et al. s (1990) transactional stress process model to study caregivers of individuals with Lewy Body Dementia, Leggett, Zarit, Taylor, and Galvin (2011) found that as motor impairment, helping with ADLs, and the care-receivers behavioral and emotional problems increased, caregivers burden and isolation did as well. In addition, using Lazarus and Folkman s (1984) stress process model, Corbeil, Quayhagen, & Quayhagen (1999) found that over a 9 month period, as behavioral impairment of the care-receiver increased, stress of the caregiver increased. This in turn resulted in increased coping methods used by the carereceiver, which were measured as the caregivers appraisal of stress, positive reappraisals, and emotional support. Despite the usefulness of the transactional process models in the study of caregiving stress, it is surprising that relatively few researchers have incorporated time use in their research. The next section reviews time use studies in the social sciences in general, the reliability and validity of time use methodology, and concludes with an overview of time use research among populations of older adults. Time Use Time use studies have been conducted for decades, with a range of populations, across a wide variety of settings. For the past five decades, the American Heritage Time Use Study has 2

8 been collecting time diary data on a range of daily experiences. Using these data, Fisher and Robinson (2010) mapped the daily routines of individuals living in twenty two different countries. The results are extremely vast and document everything from the prayer habits of Muslims (who spend more time praying on Friday), to the study habits of Bulgarians (who tend to study harder on the weekends), to music habits and pet care in different countries (people in Australia and the Netherlands spend the most time listening to music; people in France spend more time caring for pets). Time diary data have been used to study individual constructs as well. They have explored sleep (e.g., Adam, Snell, & Pendry, 2007; Chatzitheochari, & Arber, 2009), paid employment (e.g., Niemi, 2009; Sayer, 2005), and family time (e.g., Fisher & Robinson, 2010; Kalenkoski, Ribar, & Stratton, 2005). In addition, researchers have explored gender and found that there were still gender disparities in time use for labor, free time, and family interactions (e.g., Bianchi, Milkie, Sayer, & Robison, 2000; Fisher, Egerton, Gershuny, & Robinson, 2007; Mattingly & Bianchi, 2003; Mattingly & Sayer, 2006; Niemi, 2009). These studies found that men on average had more free time then women (Mattingly & Bianchi, 2003) and felt less rushed during their leisurely activities (Mattingly & Sayer, 2006). In addition, men decreased their time in the work force compared to women (Fisher, Egerton, Gershuny, & Robinson, 2007), but increased their time participating in household labor (Fisher et al., 2007; Bianchi, Milkie, Sayer, & Robison, 2000). Finally, time diaries have also explored the differences in activities among people of various ages. Robinson and Godbey (1997) compared the time use of adolescents (age 3 11), teenagers (age 12 17) and young adults (age 18 24) in various activities. They found that on average, children were able to obtain 4 6 hours more sleep than teenagers and young adults, and generally had the most free time. When comparing teenagers to young adults, the researchers found that teenagers spent 4-6 hours extra napping during the day, and performed more discretionary activities (such as socializing and watching television). In addition, using a sample of individuals age and comparing them to a sample of adults 65+, they found that even though older adults had more free time, some activities declined in importance as they aged. They were less likely to attend movies, sporting events, and other entertainment events than those younger than 65. Their leisure activities consisted mainly of a reliance on media (e.g., television, radio, and newspaper). This finding is consistent with Horgas, Wilms, and Baltes 3

9 (1998), who found that although older adults spent a large portion of their day in obligatory activities (e.g., eating, grooming), the most of their time was spent performing discretionary activities, especially watching television. Time use research serves a dual nature methodologically; it provides researchers with rich qualitative data from the words of participants, but can also be quantified (Wilcock, 1999). The most comprehensive way to collect time use data is through time diary methodology (Harvey, 1999; Harvey & Pentland, 1999). Time diaries typically retrospectively record a 24 hour log of every activity (of a certain duration) that has taken place, the physical location where the activity took place, with whom the activity was performed, the affect experienced during the activity and the duration of the activity (Harvey, 1999). To ensure that every action is accounted for in one day, the 24 hour time diaries are logged in minutes; because all 1440 minutes in a day are logged, trade-offs among activities can be reviewed (Robinson, 1999) as well as proportions of time (e.g., how long a caregiver spends by themselves or with the person with dementia, in the current context) (Harvey, 1999). The reliability and validity of time diaries have been extensively researched (e.g., Klumb & Baltes, 1999; Pentland, Harvey, Lawton, & McColl, 1999; Ujimoto, 1990). Klumb and Baltes (1999) compared the validity of the Yesterday Interview, a form of time budget methodology, to an alternative experience sampling method. They found that in terms of timing, activities participated in, and mood ratings, convergent and discriminant correlations revealed high levels of agreement between both approaches. There inter-rater reliabilities are relatively high on all measures (greater than.80), with the exception of self-care, which had a reliability of.67 (Lomax, Brown, & Howard, 2004). Pentland, Harvey, Lawton, and McColl (1999) reported reliability statistics from a number of studies in their book on time use. Robinson (1977) found a.95 correlation between time use patterns in the national time use diaries and the US-Jackson time study. Similarly high correlations were found by Harvey and Elliott (1983). Time budget methodology has been used to document time use in different environmental contexts of older adults (e.g., Baltes, Wahl, & Schmid-Furstoss, 1990; Pruchno & Rose, 2002; Moss & Lawton, 1982; Simonsick, Kasper, & Phillips, 1998). Using time budget methodology, Moss and Lawton (1982) found that for four separate groups of individuals (independent community residents, public housing tenants, recipients of intensive in-home services, and an institutional waiting list group), a larger portion of the day was dedicated to doing obligatory 4

10 activities (e.g., housework, cooking, shopping) than discretionary activities (e.g., socializing, using media) and was spent mainly at home. Specifically, they found that across the four groups, older adults spent the majority of their day watching television (20 23%) or resting (12 20%). Pruchno and Rose (2002) found similar results; they reported that frail older adults spent a large amount of their day alone, performing either discretionary or obligatory behaviors, and that a major portion of their day was also spent watching television (15%) or resting (14.5%). Time Use among Caregivers to a Relative with Dementia The failure of researchers to include time use in caregiving research that employs transactional process models of caregiving is surprising given caregiving s frequent characterization as a 24 hours a day, seven days per week responsibility (e.g., Mace & Rabins, 1999). For example, we know that assisting with/performing instrumental activities of daily living (IADLs) and activities of daily living (ADLs), as well as cueing, interacting and communicating with the person with dementia, is time consuming endeavor (e.g., Jansson, Nordberg, Garfstom, 2001; Schulz & Martire, 2004). In addition, Ory, Hoffman, Yee, Tennstedt, and Schulz (1999) reported that dementia caregivers spent significantly more hours per week providing care than non-dementia caregivers. Finally, among an investigation of 1,181 caregivers to a relative with dementia, one-third of the sample reported spending more than 14 hours providing care for their relative with dementia; one-half of those caring for a relative with severe dementia, and two-fifths of those caring for a relative with moderate dementia, reported spending at least 10 hours of their day providing care (George, Jansen, Jackson, Meyrieux, Sadowska, & Selmes, 2008). For the most part, empirical evidence about caregivers time use is relatively fragmented. Beaudreau, Spira, Gray, Depp, Long, Rothkopf, and Gallagher-Thompson (2008) found that the caregivers to individuals with dementia woke up, on average, 17.2 times a night. Because of their poor sleep schedules, they tended to suffer from more depressive symptoms. In a qualitative study, Arber and Venn (2010) found that caregivers were awake during the middle of the night from either attending to the physical needs of the care-receiver, or awake because of the anticipation of caring. This, in addition to worrying and feeling anxiety and guilt, kept the caregiver in a disrupted sleep pattern. Research exploring caregivers time spent in certain activities typically looks at one activity and its effect on caregivers health. For example, Mausbach, Harmell, Moore, and 5

11 Chattillion (2011) explored the association of leisure activities on caregiver well-being and affect. They found a strong correlation between leisure satisfactions and affect in caregivers who reported greater burden compared to those with lower burden. Caregivers who reported higher levels of burden had an increase of positive affect while participating in leisurely activities and a decrease in negative affect. Caregivers with a lower level of burden had an increase in positive affect during leisure activities, but there was no association for negative affect. Similarly, Jacobson, Dobson, Truax, Addis, Koerner, and Gollan, (1996) also found that engaging in leisurely activities decreased depressive symptomatology of caregivers for relatives with dementia. Kim, Chang, Rose and Kim (2012) explored caregiver s well-being on performing caregiver duties (i.e., help with ADLs and IADLs). They found that as time spent participating in caregiver duties, caregiver burden increased. Perhaps the most comprehensive examination of time use among caregivers to a relative with dementia, Moss, Lawton, Kleban, and Duhamel (1993) compared the family caregivers time use before and after institutionalization of their relative. They found that prior to institutionalization, 12.4% of caregivers days was dedicated to caregiving duties (e.g., medical assistance, help with ADLs and IADLs), 52.4% was spent performing other obligatory activities (e.g., shopping, housework, personal care), and 35.2% comprised discretionary activities (e.g., interacting with family and friends, religious activities, watching television). This delegation of time use to activities other than caregiving duties flies in the face of conventional characterizations of caregiving as a 36 hour day (the title of Mace and Rabins 1999 classic book on caregiving). Collectively, these studies demonstrate the valuable information gathered by time budget methodology. They offer a glimpse into the lifestyles of older adults, providing rich, contextual information about their experiences and behaviors that can be used to guide academic, planning, and policy development. Taken together, these studies do give a brief look at specific aspects of caregivers days. What has not been extensively studied however is how a caregiver to a person with dementia spends her/his time throughout the day as a whole in the context of competing responsibilities. This research overcomes this limitation by looking at caregivers use of time for a recent, complete 24 hour day and examines time use as a function of key background and contextual characteristics, stressors, and outcomes from the transactional stress process model. 6

12 The overall purpose of this research was to understand what takes place during a specific day in the life of 33 caregivers. Using the transactional process model as a heuristic, this research explored how key constructs from the model are associated with caregivers time use. Figure 1 demonstrates how the constructs being used in this analysis fit into specific components of the transactional process model. Additions to the model are in bolder black boxes. Figure 1: Modified Transactional Stress Process Model Specifically, five variables were used to look at background and contextual characteristics (i.e., gender, kinship, employment status, length of time in the caregiving role, and level of impairment of the relative with dementia), five were used to examine caregiver stressors (i.e., Hassles, Loss of Self, Role Competence, Capability, and Role Captivity) and two were used to explore caregiver well-being (i.e., burden and depressive symptomatology). Specific questions explored were: 1. In a recent 24 hour period, how did caregivers to a relative with dementia spend their time? What was their overall affect during the day? 2. How much time do caregivers actually spend with their relative with dementia during a particular day? What was their affect during these activities? 3. To what extent are background and contextual characteristics (i.e., gender, kinship, employment status, length of time in the caregiving role, and level of impairment of the relative with dementia), stressors, Loss of Self, Role Competence, Capability, Role Captivity, and 7

13 caregiver well-being (i.e., burden and depressive symptomatology) associate with caregivers overall time use? Method Participants and Recruitment Participants were recruited through the Area Office on Aging of Northwest, Ohio, Inc. and the local Alzheimer s Association chapter. A telephone screening tool developed for this research was used to determine participants eligibility and obtain information regarding their personal characteristics (e.g., age, gender, occupation), as well as information regarding the person with dementia (e.g., ADLs, needs for assistance, challenging behaviors), and their caregiving situation (e.g., how long they have been a caregiver, how many hours they spent in a day caring, and if they received formal assistance with caregiving). After completing the telephone screening tool, home visits were scheduled with consenting caregivers. The purpose of the home visit was to conduct a face-to-face interview with the caregiver and obtain information on the cognitive functioning of the individual with dementia. Researchers worked in teams; while one was conducting a face-to-face open-ended interview with the caregiver, a second researcher conducted cognitive assessments of the person with dementia and then occupied them until the caregiver s interview had been completed. Table 1 presents descriptive data on the caregivers and the care recipients, specifically variables that were documented in the telephone screening tool and face-to-face interviews. As seen in the table, the sample was predominately female and not currently employed. Of the thirteen that were employed, their average work week totaled hours (SD = 16.12). Kinship was dichotomized for analysis, but when broken down, 9 of the caregivers were wives, 4 were husbands, 14 were daughters, 4 were sons, and 2 were daughters-in-law. Not reported in the table was race and education. The caregivers were diverse with respect to race/ethnicity, with 26 White caregivers, 5 African American, 1 Hispanic, and 1 Asian. Caregivers level of education averaged years (SD = 2.39). The caregivers in the sample reported generally high well-being, as well as feeling competent in their caregiving role. However, they did report feeling less capable in their role as a caregiver role. Of the individuals with dementia, thirteen were male and twenty were female. They ranged in age from 62 to 89, with the average age being (SD = 7.45). They were also 8

14 diverse with respect to race/ethnicity, with 28 White, 4 African American, and 1 Asian. As seen in Table 1, level of cognitive impairment was measured using the Dementia Rating Scale (DRS) (Mattis, 1973; 1988); of the twenty nine who were able to participate, their average score was 71 (SD = 37.64). Table 1: Description of Caregivers and Persons with Dementia Frequency M SD Min Max Caregiver Care Recipient Age Sex Male 9 Female 24 Occupational Status Employed 13 Unemployed 9 Retired 11 Kinship Spouse 13 Child/in-law 20 Length of caregiving (in months) Hassles Reported Loss of Self Role Captivity Capability Role Captivity Burden Depressive Symptomatology DRS Total Score DRS Attention Score DRS Initiation/ Preservation Score DRS Construction Score DRS Conceptualization Score DRS Memory Score

15 Instruments Used in Data Collection Telephone Screening Tool Background and caregiving contextual variables, as well as basic demographic variables, were assessed via a telephone screening tool that was administered to potential participants prior to scheduling a home visit to obtain signed, informed consent (caregiver) and assent (person with dementia). Specific variables and how they were measured are described below. Using an open-ended format, caregivers provided information on their age, gender, relationship to the caregiver, race/ethnicity, and occupation. In addition, they provided similar information (i.e., relationship to caregiver, age, gender, race/ethnicity, education, occupation and race/ethnicity) for each household member. Caregivers also provided information on the length of time that they had been providing care and the average amount of time they spent each day providing care. Home Visit Home visits consisted of two components: a face-to-face interview with the caregiver and a cognitive assessment of the person for whom they cared. Each component is described below. Face-to-Face Interviews with Caregivers Face-to-face interviews with caregivers included quantitative and qualitative components; only the quantitative components were used in this research. The standardized instruments that were used to assess key components of the caregiving stress process are described below, followed by the measure used to assess caregivers time use. Caregiving Stressors. Caregivers appraisals of events in caregiving as stressful (i.e., daily hassles) were collected using the Caregiving Hassles and Uplifts Scale (Kinney & Stephens, 1989). The Caregiving Hassles and Uplifts Scale consist of 42 events that typically occur during caregiving that tap five domains: 1) the person with dementia s cognitive status, 2) her/his behavior problems, 3) assistance required with ADLs, 4) assistance required with IADLs, and 5) the caregivers social network. For each item, the caregiver indicates whether the item occurred during the past week of caregiving and, if the event occurred, the extent to which it was both a hassle and an uplift on a 4 point scale ranging from it wasn t to a great deal. For this 10

16 research, only the total occurrence of events that were appraised as Hassles was used. Test-retest reliability of the occurrence of hassles was.85. Potential Mediators and Moderators of the Stress Process. Caregivers completed four scales developed by Pearlin and colleagues (1990): Loss of Self, Role Competence, Caregiving Role Captivity, and Role Overload; Pearlin and Schooler s (1978) General Mastery Scale; and Schwartzer and Jerusalem s (1993) General Perceived Self-Efficacy Scale. Because of the small number of items in each scale and the potential conceptual and empirical overlap across the scales, a series of principal components factor analyses with varimax rotation were conducted on each of the six scales. Based on the results of these analyses, General Mastery and General Perceived Self-Efficacy were aggregated into one scale that reflects caregivers Capability. An additional factor analysis was done on a combination of the General Perceived Self-Efficacy scale and the General Mastery Scale. Information on the factor analyses and Cronbach s alpha reliability analyses, and final scales are presented below and in Appendix 1 (pg. 37). Loss of Self. The extent to which caregivers felt that they lost a sense of themselves during their caregiving role was assessed using Pearlin et al. s (1990) Loss of Self scale. This scale consists of two items, How much have you lost a sense of who you are? and How much have you lost an important part of yourself? The two items were rated on a four point scale ranging from 1 ( Not at all ) to 4 ( Completely ), with a higher score indicating a higher sense of loss. Factor analysis confirmed the structure of this two item scale; both factors loaded on one scale at.929. Cronbach s alpha was.82. Role Captivity. How much the caregiver felt confined to their caregiving role was measured using Pearlin et al. s (1990) Role Captivity scale. Role Captivity (α =.71 in the current sample) consists of three items (e.g., How much do you feel trapped by your relative s illness ). The items were rated on a four point scale ranging from 1 ( Not at all ) to 4 ( Very much ), with a higher score indicating higher feelings of captivity. Factor analysis confirmed the use of this three item scale, with factor loadings ranged from.756 to.841. Role Competence. Role Competence was assessed using Pearlin et al. s (1990) 4-item scale (e.g., How confident do you feel ). Each item is rated on a 4-point scale ranging from Not at all to Very much. After recoding reverse-scored items, Cronbach s alpha was.77. Analyses indicated that deletion of the item Feel that all in all, you re a good caregiver would raise the alpha to.81. Factor analysis of this scale confirmed that this particular item did not load with the 11

17 other three scale items (factor loading of.598), with factor loadings of.736,.836, and.907. Because of this, it was omitted from the scale. The factor loadings for the final three item scale ranged from.805 to.900. Role Overload. Pearlin and colleagues (1990) Role Overload scale assessed how much caregivers felt they were overwhelmed by their caregiving duties, such as feeling exhausted or having more things to do than they could handle. Each item was rated from 1 ( Not at all ) to 4 ( Completely ). An initial computation of the alpha value (.49 in the current sample) indicated that the scale was not reliable in assessing role overload. Based on the results of the factor analysis, the scale was excluded from further analyses. General Perceived Self-Efficacy. General Perceived Self-Efficacy consists of 10 items (e.g., I can solve most problems if I invest the necessary effort) that assess an individual s belief of how to cope with a variety of difficult demands in life. For each item, caregivers rated the extent to which item was most true on a 4-point scale ranging from 1 ( Not true at all ) to 4 ( Exactly true ). An initial reliability analysis yielded a satisfactory alpha level of.78 (see Appendix 1). Factor analysis of the scale confirmed the factor structure reported by Pearlin et al. (1990), with the exception that item four I am confident that I could deal efficiently with unexpected events and item ten, I can usually handle whatever comes my way, loaded on a separate factor. However, because the distributions of these two items were similar to the distribution of the other eight items and they had high item-total correlations with the other Self- Efficacy scale, these items were retained. General Mastery. Pearlin et al. s (1978) General Mastery Scale assesses an individual s sense of personal control. It consists of 7 items (e.g., I have little control over things that happen to me ). Each item is rated on a four-point scale ranging from Strongly agree to Strongly disagree. After recoding reverse-scored items for both scales, reliability analyses were conducted. The Cronbach s alpha was.56 with item 5, Sometimes I feel that I am being pushed around in life. Because of this item s low inter-item correlation and squared-multiple correlation of.11, it was omitted from the scale. A principal components factor analysis with varimax rotation confirmed that item 5 was not acceptable for this scale in this sample. A second computation of reliability, without item 5, was.66. Caregiver Capability. Based on the low Cronbach s alpha coefficients for General Perceived Self-Efficacy and General Mastery, the two scales were combined to create a measure 12

18 of caregiver capability and a factor analysis was conducted on all 17 items. Factor analysis of the 17 items revealed three factors. Fourteen factors loaded on factor 1 (factor loadings ranged from.51 to.78); items 4 and 10 from the General Perceived Self-Efficacy scale loaded on a second factor (factor loadings =.46 and.49, respectively), and item five from General Mastery loaded on a single factor at.828. This item was deleted, resulting in an overall Cronbach s alpha of.76. Higher scores indicated more capability. Caregiver Well-Being. Caregiver well-being was assessed using two measures that were frequently examined as outcomes in caregiving research: burden and depression. The Modified Burden Interview (Zarit & Zarit, 1983) is a 22 item scale that depicts how much the relative s dementia impacts the caregiver s life. Questions range from Do you feel embarrassed by your relative s behavior? to Do you feel uncomfortable having friends over because of your relative? Responses range from 0 (never) to 4 (nearly always). Individual items are scored to obtain a composite burden score, with higher scores indicating more burden; Cronbach s alpha in the current sample was.90. Depression was assessed using The Center for Epidemiologic Studies Depression Scale (CES-D; Radloff, 1977). The CES-D is a 20-item scale designed to measure depressive symptomatology in the general population. All 20 items are general statements (e.g., I felt fearful, I was happy ) with the respondent marking how many days they felt that way during the past seven days. Items are scored with the range of 0 (less than 1 day noted), 1 (1-2 days noted), 2 (3-4 days noted), or 3 (5 7 days noted). The possible range of scores is 0 to 60, with a higher score indicating more depressive symptoms. Cronbach s alpha for the current sample was.85. Time Use. As part of the face-to-face interviews, caregivers were asked to think back to 24 hours before the researchers arrived at their home, and to walk the interviewer through the past 24 hours. Caregivers were asked to indicate what activity they were doing (see Appendix 2 for a list of activities) provided it lasted five minutes or more, where they were performing it, who they were with (e.g., alone, the person with dementia, an informal, formal, or mixed crowd), and how they would characterize the activity in terms of their affect (i.e., whether they were feeling positive, neutral, or negative while performing the activity). This enabled the researchers to collect time diaries that consisted of 24 hours (1440 minutes) of data that encompassed the caregiver s entire day. 13

19 Time diaries were collected by noting the start and end time of activities completed (see Appendix 2, pg. 39, for a list of activities), the physical location of the caregiver, with whom the caregiver spends her/his time, and their overall affect. All were summed into minutes. Appendix 3 (pg. 40) represents one page of a typical time diary. This particular time diary starts at 1PM; the caregiver was out and about with her husband, stopping at their rental and shopping. During this time they were not with the person with dementia (as noted in Was the CR with you? ) and she was experiencing positive affect. The caregiver and her husband did this activity for 90 minutes until they went for lunch at 2:30. Again, they were not with the person with dementia, and they were experiencing positive affect. Lunch lasted for 45 minutes. Because each time diary provides 24 hours (1440 minutes) of data, this time diary continues in the same fashion (activity, location, who was with, and her/his affect) until the next day at 1PM, when the researchers arrive; only a portion of the diary is shown for an illustrative purpose. These raw data were then transferred into a coding sheet (Appendix 4, pg. 41) that was organized by types of activity, beginning and ending time of the activity, where the caregiver was, who she/he was with, and her/his affect. Next, the data were coded using Andrew Harvey s (1999) concept of hypercoding, a multidimensional representation of single complex activities. Using the example above, out and about would be coded as the activity errands, because even though the caregiver was doing a number of activities, she didn t specify any particular one. After noting the start and end time and calculating the number of minutes in the activity (in this case, 90), the 90 minutes would be put in the table under informal and positive (noted in the table as INFORMAL and +). This would distinguish that for 90 minutes, the caregiver was running errands with an informal person (in this case, her husband) and experienced positive affect while doing so. The same thing would happen with the 45 minutes that the couple ate. It would be coded as eating, away from the home, for 45 minutes within the informal positive chart. When all activities were accounted for and coded, the totals of each major cell heading (i.e., time, location, who they were with) were added and each totaled 1440 minutes, a 24 hour day. Assessment of Cognitive Ability of the Person with Dementia. The Dementia Rating Scale (DRS, Mattis, 1973; 1988) assesses an individual s overall cognitive functioning. In addition to an overall raw score, the DRS contains five subscales which provide information on specific abilities, including scores on attention, initiation and preservation, construction, 14

20 conceptualization, and memory. Though various diagnostic cut off scores have been used to diagnose dementia, <132 (Green, Woodward, & Green, 1995), <129 (Monsch, Bondi, Salmon, Butters, Thal, Hansen, & Klauber, 1995), the cut off scores for the DRS is <123 (Montgomery & Costa, 1983). Results The results of this study are organized into three sections. Section one provides a general overview of the caregivers time use. Table 2 represents time use as a function of who the caregiver was with, aggregated across a 24 hour / 1440 minute day. To further explore what activities were being performed and the caregivers affect, Table 3 depicts general time use broken down into activity categories (see Appendix 2 for a list of activities) and then affect. These activity and affect categories are reported regardless of who the caregiver was with and exclude the time the caregiver spent sleeping. Coding rules for this study required the activity sleeping to be coded alone and in neutral affect. By excluding sleep time from the analyses, the research captures exactly how much time was spent with the person with dementia out of all possible minutes that could be spent. This is true for all the tables in the results, with the exception of Table 2. Section two reports on caregivers general time use while with the person with dementia (and those occasions where the caregiver and her/his relative are with others). For the analysis, time spent with the person with dementia is a combination of caregivers time spent alone with their relative, and time spent with their relative and others. By coding both time spent alone with the person with dementia and time spent with the person with dementia and others, the maximum time spent can be calculated. Section two begins with Table 4, which presents the amount of time spent with the person with dementia and others in the different activity categories. In addition, it provides the overall affect of the caregiver, regardless of the activity being done. To further explore affect, Table 4 presents data on caregivers affect while with the person with dementia and others as a function of the activity being performed. To conclude the Results section, Section three examines the extent to which background and contextual characteristics, stressors, Loss of Self, Role Competence, Capability, Role Captivity and caregiver well-being associate with caregivers overall time use (further broken down by activity) with the person with dementia (and others). To better understand how 15

21 background and contextual characteristics are associated with time spent with the person with dementia, independent t-tests were conducted to compare time use as a function of gender (male vs. female) and kinship (spouse vs. child/in-law). For the latter analysis, kinship was dichotomized to represent the spouse and children and children-in-laws. A one-way ANOVA was conducted to compare whether occupational status (i.e., employed, not employed, retired) had an effect on time spent with the person with dementia and then a series of correlations were conducted to document the association between time use and the caregivers length of time in the caregiving and person with dementias level of impairment. Additional correlational analyses were conducted to examine the associations among time use and activities with the person with dementia and others on stressors, scores on Loss of Self and Role Competence, Capability, and Role Captivity and caregiver well-being (i.e., depressive symptomatology and burden). The results of these analyses are presented in terms of the components of the transactional process model (i.e., background and contextual factors, Hassles, Loss of Self, Role Competence, Capability, and Role Captivity, and burden and depressive symptomatology). Section 1: General Time Use Table 2 presents a general description of caregiver s time use as a function of who they were with, accumulated across the 24 hour /1440 minute day. Using the total time in a day (1440 minutes) as the denominator, the percentage of time spent within each group (i.e., alone, with the person with dementia and others, and with people other than the individual with dementia) was computed by dividing the number of minutes spent with each category by As can be seen in the table, during the 24 hour period recorded in the time diary, caregivers spent almost 60% of their time alone; 27.51% of which was spent awake and 31.67% of which was spent sleeping. Almost one-quarter of the day (24.43%) was spent with the person with dementia (and, in some instances, other individuals). The remaining 16.40% of their time was spent with individuals other than the individual with dementia, including a combination of informal (e.g., spouse, child) and formal (e.g., boss, doctor) network members. When caregivers were asked, In an average day, how many hours do you spend with the person with dementia, an average of hours (SD= 7.98) was reported by 28 of the informal caregivers; roughly 8 hours greater than what was found in the diaries. 16

22 Table 2: Caregivers Time Use: With Whom They Spend Time Dimension Amount of Minutes % a M SD Min Max Who With Alone Awake Sleeping PWD b (or PWD and others) Others (informal and/or formal network members) a n Based on 1440 minutes. b n Person with dementia. Caregivers spent, on average, approximately 7 ½ hours sleeping during this particular day. Table 3 provides a description of caregiver s time use when awake broken down by activity and then affect, regardless of with whom they spent the time. To calculate time excluding sleep, the adjusted denominator for activities was the sum of all activities for each individual, divided by individual activity category, and then multiplied by 100. For affect, the denominator was the sum of all time in an affect for each person, divided by an individual affect category, and then multiplied by

23 Table 3: Description of Caregivers Time Use: Activity and Affect Dimension Amount of Minutes % a M SD Min Max Activity IADLs Leisure Basic Care Monitoring Consoling Paid Employment Affect Positive Neutral Negative a n Excludes time spent sleeping. As can be seen in the table, caregivers spent approximately the same amount of time performing leisurely activities (e.g., watching television, reading, socializing) as they did performing IADLs and basic care duties. Additionally, they spent a very small portion of their day (not even 1%) monitoring their relative with dementia, and no time providing emotional support. During this time, however, caregivers reported generally high positive affect; more than one-half of the time, regardless of activity performed or who they were with, they reported experiencing positive affect. Across all activities, only one-fourth of their time use was characterized by neutral affect, and less than 10% of their day was their affect negative. 18

24 Section 2: General Time Use with the Person with Dementia Table 4 presents caregivers overall time use by activity and then affect while with their relative with dementia (and, in some instances, others). As previously reported in Table 2, Table 4 shows the overall time spent with just the person with dementia and with the person with dementia and others (i.e., in the presence of the person with dementia. Of the minutes spent with the person with dementia, caregivers spent approximately the same amount of time performing leisurely activities as they did performing IADLs and basic care duties for the person with dementia. Monitoring the person with dementia is rather low (only 2% of their awake time was dedicated to doing so). Not surprisingly, no caregiver spent time consoling their relative or with their relative while at work. 19

25 Table 4: Description of Caregivers Time Use: With Person with Dementia a Dimension Amount of Time in Minutes % b M SD Min Max Overall time spent in the presence of the person with dementia (and others) Activity Affect IADLs Leisure Basic Care Monitoring Consoling Paid Employment Positive Neutral Negative a n Includes time spent with person with dementia and others. b n Excludes sleeping time. 20

26 In addition, as can be seen in the table, when caregivers were in the presence of their relative with dementia, almost three-fourths of the time they reported positive affect, and they characterized less than 10% of the time as negative. Table 5 presents caregivers time use while with the person with dementia (and others) as a function of the activity being performed and caregivers reported affect during each activity category. For instance, of the 53.42% of the caregiver s time spent performing leisurely activities depicted in Table 4, the majority of the that time (40.59%) was spent in positive affect, and less than 15% was reported in neutral or negative. This finding is similar across all activities, with the exception of monitoring. Though very little time was spent monitoring the person with dementia, more time was reported in negative affect than in any other affect. 21

27 Table 5: Description of Caregivers Time Use: With Person with Dementia, a Activity by Affect Dimension Amount of Minutes % b M SD Min Max Activity IADLs Leisure Basic Care Monitoring Positive Neutral Negative Positive Neutral Negative Positive Neutral Negative Positive Neutral Negative a Includes time spent with person with dementia and others. b n Excludes sleeping time. 22

28 Section 3: Associations with Time Use - Background and Contextual Characteristics, Stressors, Loss of Self, Role Competence, Capability, Role Captivity, and Well-being In order to explore the association of time use on background and contextual characteristics (i.e., gender, kinship, employment status, length of time in the caregiving role, and level of impairment of the relative with dementia), stressors, Loss of Self, Role Competence, Capability, Role Captivity, and caregivers well-being (i.e., burden and depressive symptomatology) several analyses were performed. As can be seen in Table 6, male caregivers spent significantly more time with their relative with dementia than did female caregivers and spouses spent more time than children or children-in-laws. Table 6: Impact of Background and Contextual Characteristics on Time Spent with Person with Dementia Caregiver characteristic M SD t Df Male * 31 Female * = p <.05 Spouse ** 31 Child/in-law A one-way ANOVA (not shown in table) was conducted to analyze the association of time use and activity by occupational status. The result of the analysis was significant [F (2,30) = 4.335, p=.022]. Post hoc comparisons using the Bonferroni test indicated that mean scores for those who were employed (M= , SD = ) was significantly different from those who are retired (M= , SD= ) and unemployed (M= , SD= ). Caregivers who were employed spent significantly less time with their relative with dementia than those who are not working. The overall amount of time caregivers spent with their relative, as well as the amount of time spent in each activity category with the person with dementia (and others), was correlated with length of time in the caregiving role (measured in months) and the level of impairment of 23

29 the individual with dementia (i.e., total and subscale scores from the Dementia Rating Scale score). Table 7 contains the correlation matrix. The results suggest that caregivers who had spent more time in the caregiving role were caring for individuals who had lower attention scores on the Dementia Rating Scale (r=-.387, p <.05). In addition, caregivers to more impaired relatives (as indicated by total Dementia Rating Scale scores) spent more time with them overall (r = -.368; p <.05). However, when overall time use was broken down into separate activities, only time spent monitoring the individual with dementia was significant. The amount of time monitoring the person with dementia was negatively correlated with their relative s overall DRS raw score total, (r=-.449, p <.05), as well as three of the five subscales, once again suggesting that the more severe the dementia prognosis, the more likely the caregiver was to monitor the person. 24