The Gerontologist Advance Access published November 18, 2015 The Gerontologist cite as: Gerontologist, 2015, Vol. 00, No. 00, 1–10 doi:10.1093/geront/gnv148 Advance Access publication November 18, 2015

Research Article

Does Caring for a Spouse With Dementia Accelerate Cognitive Decline? Findings From the Health and Retirement Study Downloaded from http://gerontologist.oxfordjournals.org/ at Monash University on November 21, 2015

Kara Bottiggi Dassel, PhD,*,1 Dawn C. Carr, PhD,2 and Peter Vitaliano, PhD3 Gerontology Interdisciplinary Program, University of Utah, Salt Lake City. 2Stanford Center on Longevity, Stanford University, California. 3Department of Psychiatry and Behavioral Sciences, University of Washington, Seattle. 1

*Address correspondence to Kara Bottiggi Dassel, PhD, University of Utah College of Nursing, 10 S 2000 E, Salt Lake City, UT 84108. E-mail: [email protected] Received June 12, 2015; Accepted September 17, 2015

Decision Editor: Rachel Pruchno, PhD

Abstract Purpose of the Study:  The purpose of this study is to expand our recent work, which showed that spousal dementia caregivers compared to spousal nondementia caregivers experience an accelerated rate of frailty over time, by exploring cognitive health outcomes between dementia and nondementia caregivers. Design and Methods:  Using 8 biannual waves of the Health and Retirement Study data and performance on the modified Telephone Interview for Cognitive Status, we examined changes in cognitive health among surviving spousal caregivers (N = 1,255) of individuals with dementia (n = 192) and without dementia (n = 1,063), 2 waves prior and 2 waves following the death of the care recipient. Results:  Controlling for baseline health and contextual factors (e.g., frailty status, age, education), results revealed that dementia caregivers had significantly greater cognitive decline (p < .01) compared to nondementia caregivers. Relative to 2 waves prior to the death of their spouse, dementia caregivers declined by 1.77 points relative to nondementia caregivers (0.87 points) at the time their spouses’ deaths were reported and 1.89 relative to the 1.18 points at the wave following these deaths, respectively. Implications:  The findings from this study show that spousal caregivers of persons with dementia experience accelerated cognitive decline themselves compared to nondementia caregivers. These results, along with our previous study findings, suggest that this vulnerable group could benefit from early cognitive screening and psychosocial interventions designed to help dementia caregivers better maintain their cognitive and physical health during and following their intensive caregiving responsibilities. Key words:  Bereavement, Caregiver stress, Dementia

Published by Oxford University Press on behalf of the Gerontological Society of America 2015.

1

2

levels of stress and reap many benefits from caregiving (for a review, see Roth, Fredman, & Haley, 2015). Importantly, the degree to which caregiver outcomes are negative, neutral and/ or positive may be explained by models that have focused on individual differences. For example, one model (Vitaliano, Russo, Young, Teri, & Maiuro, 1991) has emphasized the importance of vulnerabilities (gender, age, psychiatric history, disposition, genetics, early environment) and resources (social supports, coping, income) that either directly influence caregiver outcomes or interact with caregiver exposure to influence caregiver outcomes. Caregivers with fewer vulnerabilities and more resources may be more likely to experience positive benefits from caregiving. This model was supported by results showing that caregivers without a history of depression prior to caregiving have similar levels of current clinical depression (5%) to noncaregivers without a history of depression (3%). In contrast, 73% of caregivers with a history of depression have current depression compared to 30% in noncaregivers with history of depression (Russo, Vitaliano, Brewer, Katon, & Becker, 1995). This suggests that many caregivers are not at risk for depression. Similarly, higher depression in caregivers compared to noncaregivers has been shown to be largely influenced by genes and early family environment (Vitaliano, Strachan, Dansie, Goldberg, & Buchwald, 2014). Clearly, individual differences play a major role in how persons respond to caregiving. In a recent study, we examined spousal dementia versus nondementia caregivers on physical health outcomes and found that dementia caregivers were more likely to experience accelerated frailty relative to nondementia caregivers (Dassel & Carr, 2014). Our primary objective in this paper is to expand our previous findings and build on those of Vitaliano and colleagues (2011). To do this, we will examine whether dementia compared to nondementia spousal caregiving is related to cognitive health. Importantly, we examine these relationships both prior to and following the cessation of caregiving responsibilities (i.e., death of care recipient) while adjusting for selection effects (i.e., socioeconomic status) related to dementia caregiving. Guided by a proposed chronic stress model (Vitaliano et al., 2011) and our recent research that demonstrated accelerated onset of frailty, we proposed the following relationships: Hypothesis #1: Caregivers of individuals with dementia relative to caregivers of individuals without dementia will have greater declines in cognitive health two waves prior (wave(t − 2)) to the death of the care recipient (wave(t)). Hypothesis #2:  Cognitive health decline in caregivers of care recipients with dementia will be greater than in caregivers of care recipients without dementia two waves prior (wave(t − 2)) to one wave following (wave(t ) the cessation of caregiving responsi+ 1) bilities (i.e., death of care recipient).

Downloaded from http://gerontologist.oxfordjournals.org/ at Monash University on November 21, 2015

Dementia caregivers are at greater risk for negative health outcomes compared to demographically similar noncaregivers (Braun et al., 2009; Vitaliano, Zhang, & Scanlan, 2003). They are also more likely to experience depression (Pinquart & Sorenson, 2003), social isolation (Ory, Hoffmann, & Yee, 1999), obesity (Vitaliano, Murphy, Young, Echeverria, & Borson, 2011), and increased levels of cortisol and reduced antibodies (Vitaliano et  al., 2003). Of particular interest to the current study is the fact that caregivers may experience increased risk of subsequent dementia diagnosis themselves (Norton et al., 2010) or cognitive problems or decline relative to noncaregivers (de Vugt et al., 2006; Vitaliano et al., 2011). The impact of short- and long-term exposure to stress on cognitive abilities in caregivers has been shown to be related to deficits in a variety of cognitive domains including immediate, delayed, and working memory (Lee, Kawachi, & Grodstein, 2004; Mackenzie, Wiprzycka, Hasher, & Goldstein, 2009; McEwen, 2000), global cognitive functions (de Vugt et  al., 2006), psychomotor speed (Vitaliano, Echeverria, Shelkey, Zhang, & Scanlan, 2007), and executive functions (MacKenzie, Smith, Hasher, Leach, & Behl, 2007). Older adult caregivers may be particularly vulnerable to the physiological effects of stress on cognitive status due to an age-associated overresponsiveness of the hypothalamic–pituitary axis (HPA) to stressors combined with reduced HPA sensitivity to glucocorticoid feedback (McEwen, 2000). A  meta-analysis by Vitaliano and colleagues (2003) examined the physiological effects of chronic stress between dementia caregivers and noncaregivers who were similar in age and gender. They found that dementia caregivers were significantly more likely to have elevated levels of stress hormones (i.e., glucocorticoids) and lower levels of antibody response compared to noncaregivers. Vitaliano and colleagues (2011), in their chronic stress model, proposed that spousal dementia caregivers were at additional risk for cognitive problems than demographically similar noncaregivers because of psychosocial(e.g., depression, poor sleep, social isolation), behavioral (e.g., sedentary lifestyle, poor diet), and physiological factors (e.g., increased inflammation and metabolic abnormalities) associated with caring for a loved one with dementia. These could also place spousal dementia caregivers at increased risk for cognitive decline themselves—a phenomenon the authors referred to as an “ironic tragedy” (p. 900). Additional factors that place spousal dementia caregivers at risk for cognitive decline include their advanced age (dementia caregivers are typically over the age of 60) and the associated increased risk of dementia (Kawas, Gray, Brookmeyer, Fozard, & Zonderman, 2000) as well as additional stress associated with the fact that spouses typically provide significantly more hours of care per week (+35 on average) compared to nonspousal caregivers (e.g., adult children) (Ory et al., 1999). Although numerous studies have suggested that caregiving is stressful and that it is associated with psychological distress and physiological dysregulation, more recent population studies have shown that many caregivers report low

The Gerontologist, 2015, Vol. 00, No. 00

The Gerontologist, 2015, Vol. 00, No. 00

Design and Methods Data and Sample

we included those with data available for the duration of the study period (i.e., with cognitive data at each 2-year interval from wave(t − 2) to wave(t + 1)). Our total sample included 1,255 surviving spouses, among whom 192 (14.8%) cared for a spouse with dementia.

Change in Cognition Our previous work has shown that spousal caregivers of persons with dementia decline on Digit Symbol Test over 2  years relative to matched noncaregivers, with an effect size of (d = 0.29) corresponding to a decline of 3.3 points (Vitaliano et  al., 2009). Also, in the same period, these same caregivers showed a small (decline from 35.5 points to 34.5 points, d  =  0.33), but significant decline on the Shipley Vocabulary Test relative to matched noncaregivers, suggesting a trend towards a decline in verbal intelligence (i.e., dementia) (Vitaliano, Echeverria, Yi, Phillips, Young, & Siegler, 2005). To our knowledge, this is the first to explore changes in cognitive health between caregivers whose spouses died of dementia and caregivers whose spouses did not die of dementia. To examine cognitive changes over time, we used the HRS’ global cognitive performance measure. Cognitive performance of the surviving spouse was based on a 27-point measure, based on cognitive tests performed in the 1996–2010 biannual waves of the HRS. This test is derived from the Telephone Interview for Cognitive Status (TICS) (Brandt, Spencer, & Folstein, 1988). The modified TICS (mTICS) has been validated for use as a screening instrument for cognitive performance (Plassman, Newman, Welsh, Helms, & Breitner, 1994; Welsh, Breitner, & Magruder-Habib, 1993) and is a composite measure of cognition based on three aspects of cognitive function: (a) episodic memory (10-item immediate and delayed recall test; range 0–20 points); (b) working memory (serial 7-s test; range 0–5 points); and (c) processing speed (backwards-counting; range 0–2 points). Cut-points have been validated for differentiating normal cognitive functioning from impaired functioning. Scores of less than 7 (out of a possible 27 points; lower scores indicate greater impairment) on the mTICS indicates the presence of dementia in the HRS study (Crimmins, Kim, Langa, & Weier, 2011; Fisher, Rodgers, & Weir, 2009; Langa et  al., 2005). As a result, we limit our sample to only individuals who scored at least 7 at wave(t − 2).

Spousal Caregiver Types The dementia status of the decedent is based on a response in the Exit interview indicating whether, at any point prior to death, the decedent was told by a physician that she/ he had received a diagnosis of a memory-related disease. For the purposes of both our previous and the current studies, we refer to memory-related disease as “dementia” (see Dassel & Carr, 2014 for more detail). To test our

Downloaded from http://gerontologist.oxfordjournals.org/ at Monash University on November 21, 2015

To test our hypotheses, we used data from the Health and Retirement Study (HRS), a nationally representative study of adults over age 50. The HRS has collected data biannually since 1992. We used eight waves of core data (1996– 2010), which come from the user-friendly RAND HRS data files (RAND Center for the Study of Aging, 2008) and four waves of Exit data (2002–2008) to identify our sample (Health and Retirement Study, 2010; National Institute on Aging, 2007). First we tested our hypothesis that caring for a spouse with dementia is related to the cognitive health of the surviving spouse more than it does the cognitive health of someone caring for a spouse without dementia. To do this, we examined changes in the cognitive health of primary spousal caregivers for two types of surviving spouses, those whose spouse died with a diagnosed memory disease (i.e., dementia) and those whose spouse did not have a diagnosis of dementia prior to death. For the purposes of the current study, all surviving spouses who self-identified as caregivers and who completed an Exit interview following the death of their spouse and completed core surveys, were included in the sample of spousal caregivers. The caregiving variables were only included in the Exit survey and these were limited to hours of care (ranging from 0 to 24 hr) provided on a typical day, 3 months prior to death, and the types of impairments and care required. Previous studies indicate that cognitive performance is heavily impacted by the stressful circumstances surrounding the dying process (Mackenzie, Smith, Hasher, Leach, & Behl, 2007; Uglanova & Staudinger, 2013). Thus, to capture a more accurate shift in changes in cognitive health that may have occurred from caregiving and bereavement, we used a baseline cognitive score that occurred more than one full year prior to the death of the care recipient. Given the structure of the HRS, this baseline measure was two biannual study waves prior to the wave in which the death was first reported (i.e., wave(t − 2)). Hence, change in caregiver cognition was examined from wave(t − 2) to wave(t) (approximately 4 years) and wave(t − 2) to wave(t + 1) (approximately 6  years) to examine the longer term effect of the death on the cognitive health of the surviving spouse. In short, the four points of data collection were approximately 4 years prior to death (i.e., t − 2), 2 years prior to death (i.e., t − 1), within a few months following the death (i.e., t), and approximately 2 years following the death of the care recipient (i.e., t + 1). Thus, to select our sample, we identified all individuals who were married at the time of death and had a completed proxy Exit survey at the wave in which the death was reported. According to the wave at which the death was first identified, we merged data files for the surviving spouses with the coordinating Exit file, forming a pooled sample of all surviving spouses. Among these individuals,

3

The Gerontologist, 2015, Vol. 00, No. 00

4

hypotheses, we examined differences between those who cared for a spouse with and without a diagnosis of dementia. First, we assessed the change in cognitive performance from two waves prior to the death (wave(t − 2)) to the wave in which the death of the care recipient was reported (i.e., Exit wave or wave(t)). Second, we examined the change from two waves prior to the death and the wave following the death of the care recipient (wave(t + 1)).

Covariates

Analysis Plan We used several strategies to adjust for the potential selection bias related to a given “treatment” (i.e., spousal caregiving group) and cognitive performance. Most importantly, we excluded all surviving spouses whose cognitive score suggested dementia was likely at wave(t − 2) (i.e., mTICS score less than 7).

Results Descriptive Results Table 1 provides a description of the unadjusted covariates used in our models for the final study sample of 192 dementia and 1,063 nondementia spousal caregivers. Importantly, at the two waves prior to the death of the care recipient, dementia caregivers were frailer (a score of 0.19 relative to 0.14 at wave(t − 2) and 0.22 vs. 0.14 at wave(t − 1)) and older (76.4 vs. 72.4 at wave(t − 2)). Moreover, their spouses were less likely to have died at home (22% vs. 35%) and more likely to have died in a nursing home (41% vs. 11%). In contrast, no differences were observed with regard to the expectedness of the death, gender, race, education, wealth, income, or the time spent providing daily care during the last 3 months of life. Since it is well known that depression is common in dementia caregivers and has been found to be related to poor cognitive functioning (Vitaliano et  al., 2011), we examined differences in depression between groups at each wave. The only difference in depression rates was found at the wave prior to death (i.e., t − 1). In that wave, dementia caregivers had 2.27 symptoms, and non-den caregivers had 1.7 symptoms out of 8 total. By the time the death occurred and in the wave following, the depressive symptoms for both caregiving groups were not statistically different. This suggests that dementia caregivers’ depressive symptoms may have resolved more quickly than nondementia caregivers following death, but prior to death they may had experienced more symptoms due to the process of anticipatory grief. This is a common experience in dementia caregivers and is likely related to caregiver burden (Holley & Mast, 2009). There were no significant differences in scores at other waves and, in general, scores for both groups were low on average. Specifically, at baseline (wave(t − 2)) dementia had 1.76 symptoms and nondementia caregivers only had 1.48 on average. At wave(t) (i.e., the time the death was reported), they had 2.63 and 2.60, and in wave(t + 1) (the wave following the death), 1.69 and 1.83 respectively.

Downloaded from http://gerontologist.oxfordjournals.org/ at Monash University on November 21, 2015

We controlled for several pre-widowhood characteristics of the surviving spouse in wave(t − 2) shown to impact cognitive performance: sex, race (a dichotomous measure of being White relative to non-White), education (in years), age (in continuous years), wealth (in deciles), income (in deciles), and frailty. Following Yang and Lee’s (2010) index of 30 items, frailty is a single, comprehensive measure of health using: 8 chronic illnesses, 5 activities of daily living limitations, 7 instrumental activities of daily living limitations, 8 depressive symptoms (Radloff, 1977), obesity (i.e., body mass index of 30 or greater), and self-rated health. Given the great importance that these variables have for cognition (especially frailty), support for our hypotheses should make our results especially conservative. We also controlled for the circumstances surrounding the caregiving and death as described in the Exit interview. First, we included a measure of the average number hours the surviving spouse provided care during the final 3 months of life (±5 hr per day). Second, we included a dichotomous measure of whether or not the death occurred at home or in a nursing home. Finally, we included a dichotomous measure of whether or not the death was sudden or expected at the time it occurred. Our sample of 1,255 excluded 133 individuals who had complete data on all measures except the post-death survey (wave(t + 1)) cognitive performance measure and were, therefore, excluded from the study. Among these individuals, 31 did not complete the post-death survey cognitive performance measure because their survey was completed by a proxy. Among the remaining, 29 did not complete the survey for that wave, and 73 died. Using logistic regression, we determine that relative to those to remain in the study, the 133 without post-death cognitive performance data had lower pre-death cognitive performance scores, were older, and were more likely to be male but did not differ from the selected sample based on other characteristics.

In order to address or primary research questions, we first examined bivariate differences between dementia groups (n = 192) and nondementia caregivers (n = 1,063) at all data points (see Table  1). Importantly, we included cognition and frailty measures at both wave(t − 2) and wave(t in all analyses in order to capture the pre-death changes − 1) that may have been influential during the caregiving period. Second, ordinary least squares (OLS) regression was used to assess the relationship of the decedent’s cognitive status relative to declines in cognitive health of the caregiver between wave(t − 2) and wave(t) and between wave(t) and at wave(t + 1). Caregiver demographic, care responsibilities, and health covariates described above were also included in the OLS models (see Table 2).

The Gerontologist, 2015, Vol. 00, No. 00

5

Table 1.  Descriptive Statistics: Total and Group Caregiver Differences in Cognitive, Demographic, and Health Characteristics Total sample (N = 1,255)

SD

15.731 15.282 14.722 14.441

4.039 4.062 4.336 4.490

0.153

Min.

Nondementia CG (n = 1,063)

Max.

Mean

SD

Mean

SD

7 2 2 0

27 27 27 27

16.104 15.125 14.349 14.172

4.017 3.927 4.219 4.646

15.663 15.310 14.790 14.490

4.042 4.087 4.355 4.463

0.360

0

1

−1.008 −1.289

3.810 3.988

−18 −19

17 16

0.324 0.158 0.331

0.468 0.365 0.471

0 0 0

1 1 1

15.282 0.738 0.166 0.186 0.795 12.024 73.000 5.555 0.367

4.062 0.440 0.132 0.139 0.404 3.051 9.668 2.745 0.482

2 0 0 0 0 0 40 1 0

27 1 0.805 0.873 1 17 96 10 1

1 −1.755** −1.932*

0

0

0

3.751 4.162

−0.873 −1.173

3.806 3.947

0.297 0.411*** 0.219***

0.458 0.493 0.414

0.328 0.112 0.352

0.470 0.315 0.478

15.125 0.714 0.191* 0.218* 0.807 12.214 76.417*** 5.604 0.385

3.927 0.453 0.138 0.137 0.395 3.252 8.933 2.847 0.488

15.310 0.742 0.161 0.180 0.793 11.990 72.383 5.547 0.364

4.087 0.438 0.130 0.138 0.405 3.014 9.671 2.728 0.481

Note: CG, caregiver. *p < .05. **p < .01. ***p < .001; significance indicates differences between dementia and nondementia caregivers using chi-square or t test difference tests.

To test our hypotheses, that caregivers of individuals with dementia relative to caregivers of individuals without dementia would have greater cognitive decline over time (both prior to and following the death of the care recipient), we began by examining the average cognitive scores over time for dementia and nondementia caregivers. There were no significant differences in cognitive test scores on the mTICS between caregiver groups at both wave(t − 2) (16.10 vs. 15.66 points) and wave(t − 1) (15.13 vs. 15.31 points). In contrast, the dementia caregivers performing significantly less well on the mTICS compared to the nondementia caregivers between wave(t − 2) and wave(t) (−1.76 vs. −0.87 point decline) and between wave(t − 2) and wave(t + 1) (−1.93 vs. −1.17 point decline; see Figure 1).

Linear Regression Models Results Table 2 shows the results of the two OLS regression models for changes in caregiver cognitive health between wave(t − 2) and wave(t) and between wave(t − 2) and wave(t + 1). Results for change in cognitive performance (i.e., mTICS score) between wave(t − 2) and wave(t) indicate that, when controlling for all other factors, being a dementia

spousal caregiver versus a nondementia caregiver is associated with an estimated beta of −0.87 (p < .01). Results for the change in cognitive status between wave(t − 2) and wave(t + 1) indicate that, when controlling for all other factors, being a dementia spousal caregiver versus a nondementia spousal caregiver is associated with an estimated beta of −0.75, (p < .01). These two results suggest that the cognitive difference between dementia and nondementia caregivers persists in the 2  years following the reported deaths. The location of the death of the care recipient (i.e., home vs. nursing home) and age of the caregiver were each significant predictors (p < .001) of changes in caregiver cognitive health between wave(t − 2) and wave(t + 1). To further examine the differences between dementia and nondementia caregivers with regard to cognitive performance changes, we calculated the predicted probabilities of changes in cognitive performance for dementia versus nondementia caregivers when calculating all other factors in the full regression models as the mean. Results are shown in Figure 1. The estimate for decline in the mTICS for dementia caregivers between wave(t − 2) and wave(t) is a decline of −1.77 points and for nondementia caregivers,

Downloaded from http://gerontologist.oxfordjournals.org/ at Monash University on November 21, 2015

Cognitive performance scores over time   Cognition (wave(t − 2))   Cognition (wave(t − 1))   Cognition (wave(t))   Cognition (wave(t + 1)) Variables used in regression models   Primary independent variable    Spouse with dementia   Outcome measures    Change in cognition wave(t – 2) to wave(t)    Change in cognition wave(t − 2) to wave(t + 1)   Control variables    Care recipient factors    Death sudden     Died in a nursing home    Died at home    Care receiver characteristics    Cognition (wave(t − 1))    Female    Frailty index (wave(t − 2))    Frailty index (wave(t − 1))    Non-Hispanic White    Education (years)    Age    Wealth    5+ hr/day care provided

Mean

Dementia CG (n = 192)

The Gerontologist, 2015, Vol. 00, No. 00

6

Table 2.  Marginal Effects Model Results of Change Scores Across Waves (N = 1,255)

Primary independent variable   Dementia status

Change in cognition wave(t − 2) to wave(t)

Change in cognition wave(t − 2) to wave(t + 1)

β

β

β

β

−0.90** 0.32

−0.75* 0.31

−0.71* 0.33

−0.87** 0.30

Care recipient factors   Death sudden

0.07 0.23 0.39 0.33 −0.22 0.24

  Died in a nursing home   Died at home

0.05* 0.03

 Female   Frailty index (wave(t − 2))   Frailty index (wave(t − 1))   Non-Hispanic White   Education (years)  Age  Wealth   5+ hr/day care provided  Intercept   L2 (model)

−1.69*** 0.42 −3452.5

0.04 0.03 −0.19 0.25 −0.21 1.36 −0.11 1.27 0.14 0.30 0.00 0.04 −0.03* 0.01 0.01 0.05 0.19 0.23 0.81 1.14 −3,448

0.06* 0.03

−2.16*** 0.44 −3,511

0.04 0.03 −0.32 0.26 −0.35 1.41 0.71 1.33 −0.15 0.32 0.00 0.04 −0.06*** 0.01 0.06 0.05 0.25 0.24 2.14 1.19 −3,499

Note: *p < .05. **p < .01. ***p < .001.

Figure 1.  Difference between caregiver groups on cognitive test scores (points) on the mTICS (modified Telephone Interview for Cognitive Status).

a decline of −0.87 points. Between wave(t − 2) and wave(t + , the estimate for decline in the mTICS was −1.89 points 1) for dementia caregivers and −1.18 points for nondementia caregivers.

Discussion In this study, we used HRS data to examine differential cognitive changes over three and over four biannual study waves in spouses caring for persons with or without a

Downloaded from http://gerontologist.oxfordjournals.org/ at Monash University on November 21, 2015

Caregiver factors   Cognition (wave(t − 1))

−0.15 0.24 0.41 0.34 −0.18 0.25

The Gerontologist, 2015, Vol. 00, No. 00

Location of death (i.e., died in a nursing home) was also a significant predictor of cognitive change scores over time in spousal caregivers. Not surprisingly, significantly more care recipients with dementia died in a nursing home compared to care recipients who were cognitively intact (41% vs. 11%). Research has shown that in response to increasing dementia severity, caregiver burden, and behavioral and nighttime disturbances, as many as 90% of persons with Alzheimer’s disease are institutionalized prior to death (Smith, Kokmen, & O’Brien, 2000; Yaffe et  al., 2002). However, even when spousal caregivers relinquish formal care by placing their loved one in a residential facility, many still remain highly engaged in physical care activities and report high levels of anxiety and depression (Gaugler, Mittleman, Hepburn, & Newcomer, 2010; Schulz et  al., 2004). The current findings are consistent with our previous study (Dassel & Carr, 2014), which found that, in this same group of caregivers, dementia caregivers had an accelerated rate of frailty over time compared to nondementia spousal caregivers. The results of the current study extend the previous findings of accelerated physical frailty to a similar decline in cognitive health in dementia caregivers. The combined results and the chronic stress model (Vitaliano et  al., 2011) suggest that dementia caregivers may experience greater cognitive losses than their nondementia caregiving counterparts even while controlling for physical frailty. Furthermore, the differences in cognitive health extend, on average, 2 years following the cessation of care responsibilities. It is important to note that our results may underestimate the degree of cognitive decline experienced by the general population of dementia spousal caregivers. This may have occurred for three reasons. We excluded caregivers who were (a) older, (b) had lower cognitive performance), and (c) were frail. Moreover, the frailty measure included depression, which has been shown to mediate the relationship between caregiver status and processing speed (Vitaliano et  al., 2009). Hence, by controlling for depression, we may have reduced some of the influence of dementia caregiving on cognition because we showed that dementia caregivers have greater increases in frailty over time than do nondementia caregivers (Dassel & Carr, 2014). Our study is not without limitations. First, the HRS does not ask spouses to self-identify as a caregiver for their spouses and, further, are not assessed with respect to caregiving stressors in the HRS. This requires us to make the assumption that, by virtue of being married to someone who is dying, a spouse is providing at least some amount of care. However, we do not know the intensity of this care or the stress it imposes. These factors would be beneficial in examining group differences and identifying particular areas related to the effects of spousal caregiving at the end of life that are particularly important with respect to cognitive status. Second, our primary cognitive performance

Downloaded from http://gerontologist.oxfordjournals.org/ at Monash University on November 21, 2015

diagnosis of dementia. The results support our two hypotheses that dementia caregivers would experience poorer cognitive health: (a) two waves prior to the death of the care recipient and (b) one wave following the death of the care recipient. These results are significant for two reasons. First, they support previous research that dementia caregivers are vulnerable to cognitive decline during the caregiving period. However, they also suggest that this decline is observed following the cessation of caregiving responsibilities (i.e., death of care recipient). This is very important because these results suggest that dementia spousal caregivers experience cognitive decline that is not merely transient from caregiver factors such as caregiving responsibilities and distress but may also be from an underlying neuropathological disease process. One could argue that small change scores of close to 1 point on a 27-point measure of global cognition may not be clinically significant (i.e., unless reaching the cutoff for dementia), however, the consistent declines observed in both groups over time may have an impact on practical functioning in this vulnerable population. Furthermore, the average mTICS scores of both caregiver groups were midrange with an average baseline score 16, which places this group closer to scores that indicate significant cognitive impairment (i.e., cutoff scores of 7), so a 1-point loss could have significant clinical and diagnostic implications. Indeed, small cognitive changes in dementia caregivers relative to demographically similar noncaregivers have been shown to be mediated by psychophysiological measures known to be associated with cognition (Vitaliano et al., 2005). One paper postulated a model to potentially explain greater levels of cognitive problems in caregivers relative to noncaregivers (Vitaliano et al., 2011). It suggested a priori, moderating, and mediating factors to potentially explain cognitive problems in caregivers relative to noncaregivers. In the current paper, the modeling was fairly complex, and although it did not include measures of several potential mechanisms, it did include demographic measures, depression and frailty. The last is notable as it is related to many factors associated with caregiver vulnerabilities. Two other factors were significant predictors of changes in cognitive health between caregiver groups in the extended time frame of two waves prior through 1  year following the death of the care recipient—location of death and age of the caregiver. The dementia caregiver group was significantly older than the nondementia caregivers, thereby, making them more likely to experience greater cognitive declines over time compared to their younger caregiver counterparts. It is well known that increasing age is associated with increasing rates of Alzheimer’s disease and related dementias (Kawas et al., 2000; Lindsay et al., 2002). However, although age was a significant predictor of cognitive decline in caregivers, care recipient cognitive status remained an independent significant predictor suggesting that the cognitive differences are not solely attributable to caregiver age differences.

7

The Gerontologist, 2015, Vol. 00, No. 00

8

compared to nondementia caregivers. Our findings have significant health, financial, and policy implications. With the growing older adult population, increased prevalence of dementia (Hebert, Weuve, Scherr, & Evans, 2013) and subsequent caregiving demands placed on spousal caregivers, it is imperative that psychosocial therapeutic and educational interventions be targeted to factors that will not only reduce stress but also improve health outcomes and quality of life in both caregivers and individuals with dementia (Gallagher-Thompson & Coon, 2007; Garand et al., 2014; Mittelman, Roth, Clay, & Haley, 2007; Olazarán et  al., 2010). Further research in this area will help identify underlying mechanisms (e.g., pre-caregiving [shared environment, genetics, lifetime history of illnesses] and current vulnerabilities [resources, shared environment, resilience, lifestyle factors]) that may make some spousal dementia caregivers at greater risk for cognitive decline. Specifically, recent studies have examined the issue of assortative mating or homogamy in relation to cognitive performance in spousal dyads and have found high correlations on cognitive abilities (Gualtieri, 2013; Vitaliano et  al., 2011). Future research that accounts for vulnerabilities (e.g., genetics, early environment, demographics, personality), current morbidities, life stressors, resources (social supports, coping, income), lifestyle factors, and shared pre- and postcaregiver environment will go a long way to better explaining which caregivers are at risk for cognitive decline. Moreover, additional research regarding the unique bereavement process (i.e., ambiguous loss and anticipatory grief) will aid in the development of depression-focused caregiver intervention programs (Haley et al., 2008). According to the 2015  “Facts and Figures” report produced by the Alzheimer’s Association, approximately two third of persons providing care for a loved one with Alzheimer’s disease are women and 34% of women caregivers are aged 65 or older (Alzheimer’s Association, 2015). This is a particular concern because older women are a group that already has higher risk of developing Alzheimer’s disease without the additional stressors of being a dementia caregiver. Financially, informal caregivers, including spouses, provide approximately 17.9 billion hours of unpaid care, which is estimated to value $217.7 billion annually. In addition, the negative health outcomes associated with caregiving have been estimated to add an additional $9.7 billion annually. The health and cognition of spousal caregivers is not only imperative to ensuring the best health outcomes of the care recipient, but also to avoid a progressive decline in cognitive and/or physiological health of the spousal caregiver themselves.

References Alzheimer’s Association. (2015). 2015 Alzheimer’s disease facts and figures. Alzheimer’s & Dementia, 11, 332–384.

Downloaded from http://gerontologist.oxfordjournals.org/ at Monash University on November 21, 2015

scale (i.e., mTICS) is a brief global cognition measure that primarily measures memory and processing speed, and thereby does not provide an in-depth examination of how stress impacts other cognitive domains such as language and visuospatial functions. Third, we have limited data on the surviving spouses’ transition into a caregiving role prior to the Exit interview. This interview, which occurred during the study wave in which the death was first reported, does not inquire about specific caregiving responsibilities of the spouse and, rather, focus on the care provided during the 3  months prior to the death. However, Norton and coworkers (2010) found that spouses of persons with dementia had approximately a one and a half time greater risk of incident dementia in the absence of data regarding identification as a “caregiver” or information on care responsibilities or perceived stress—a finding consistent with the current study. Furthermore, the entry into a caregiving role may be subtle and difficult for caregivers to identify (Pearlin & Aneshensel, 1994; Seltzer & Li, 1996) and based on the “hierarchical compensatory” model (Cantor, 1975 as cited in Gaugler, Wackerbarth, Mendiondo, Schmitt, & Smith, 2003), care is provided by the most accessible person (i.e., typically the spouse followed by adult children). Despite these limitations, this study has several strengths. First, unlike much of the previous research that examined relationships of dementia caregiving and cognitive function/ decline, we controlled for caregiver demographic, health, and care responsibility variables thereby controlling for several potential confounding effects on cognition that could result from the selection bias of being a dementia caregiver. Second, to our knowledge, this is the first study to explore cognitive health outcomes between two caregiver spousal groups who experienced widowhood. Our longitudinal design spanned approximately 8  years and not only captured cognitive status during the caregiving period, but also through the bereavement process when the residual stress of the caregiving experience might continue to impact the surviving spouse. Lastly, these findings expand our previous findings regarding physical health outcomes in a similar study population, thereby providing a more holistic picture of how dementia caregiving is related to cognitive health. Future studies should consider adding physiological measures of stress in the assessment of caregiver cognitive health outcomes in order to elucidate the relationships among these variables. For example, both general literature and caregiver literature (Vitaliano et al., 2005) suggest that metabolic factors may predict greater cognitive decline in the general population and mediate greater cognitive decline in caregivers versus noncaregivers. There is also a need to examine the influence of the genetic risk factor, apolipoprotein e4 variant, as a contributor to cognitive and physical vulnerability over time in this particular population of spousal caregivers. In summary, based on our previous and current findings, spousal dementia caregivers are more likely to experience declines in physical and/or cognitive health over time

The Gerontologist, 2015, Vol. 00, No. 00

Kawas, C., Gray, S., Brookmeyer, R., Fozard, J., & Zonderman, A. (2000). Age-specific incidence rates of Alzheimer’s disease: The Baltimore Longitudinal Study of Aging. Neurology, 54, 2072– 2077. doi:10.1212/WNL.54.11.2072 Langa, K. M., Plassman, B. L., Wallace, R. B., Herzog, A. R., Heerings, S. G., Ofstedal, M. B., … Wilis, R. J. (2005). The Aging, Demographics, and Memory Study: Study design and methods. Neuroepidemiology, 25, 181–191. Lee, S., Kawachi, I., & Grodstein, F. (2004). Does caregiving stress affect cognitive function in older women? Journal of Nervous and Mental Disorders, 192, 51–57. Lindsay, J., Laurin, D., Verreault, R., Hebert, R., Hellwell, B., Hill, G. B., & McDowell, I. (2002) Risk factors for Alzheimer’s disease: A prospective analysis from the Canadian Study of Health and Aging. American Journal of Epidemiology, 156, 445–453. doi:10.1093/aje/kwf074 Mackenzie, C. S., Smith, M. C., Hasher, L., Leach, L., & Behl, P. (2007). Cognitive functioning under stress: Evidence from the informal caregivers of palliative patients. Journal of Palliative Medicine, 10, 749–758. Mackenzie, C. S., Wiprzycka, U. J., Hasher, L., & Goldstein, D. (2009). Associations between psychological distress, learning, and memory in spouse caregivers of older adults. Journals of Gerontology Series B: Psychological Science and Social Science, 64, 742–746. doi:10.1093/geronb/gbp076 McEwen, B. S. (2000). The neurobiology of stress: From serendipity to clinical relevance. Brain Research, 886, 1432–1434. Mittelman, M. S., Roth, D. L., Clay, O. J., & Haley, W. E. (2007). Preserving health of Alzheimer caregivers: Impact of a spouse caregiver intervention. American Journal of Geriatric Psychiatry, 15, 780–789. National Institute on Aging. 2007. The Health & Retirement Study: Growing older in America. NIH Publication No. 07-5757. Norton, M. C., Smith, K. R., Ostbye, T., Tschanz, J. T., Corcoran, C., Schwartz, S., … Cache County Investigators. (2010). Greater risk of dementia when spouse has dementia? The Cache County Study, Journal of the American Geriatrics Society, 58, 895–900. doi:10.1111/j.1532-5415.2010.02806.x. Olazarán, J., Reisberg, B., Clare, L., Cruz, I., Peña-Casanova, J., Del Ser, T., … Muñiz, R. (2010). Nonpharmacological therapies in Alzheimer's disease: A  systematic review of efficacy. Dementia and Geriatric Cognitive Disorders, 30, 161–178. doi:10.1159/000316119 Ory, M. G., Hoffmann, R. R., III, Yee, J. L., Tennstedt, S., & Schulz, R. (1999). Prevalence and impact of caregiving: A  detailed comparison between dementia and nondementia caregivers. Gerontologist, 39, 177–185. doi:10.1093/geront/39.2.177 Pearlin, L. I., & Aneshensel, C. S. (1994). Caregiving: The unexpected career. Social Justice Research, 7, 373–390. Pinquart, M., & Sorenson, S. (2003). Difference s between caregivers and noncaregivers in psychological and physical health: A metaanalysis. Psychology of Aging, 18, 250–267. Plassman, B. L., Newman, T., Welsh, K. A., Helms, M., Breitner, J. C.  S. (1994). Application in epidemiological and longitudinal studies. Neuropsychiatry, Neuropsychology, & Behavioral Neurology, 7, 235–241. Radloff, L. S. (1977). The CES-D scale: A self-report depression scale for research in the general population. Applied Psychological Measurement, 1, 385–401.

Downloaded from http://gerontologist.oxfordjournals.org/ at Monash University on November 21, 2015

Brandt, J., Spencer, M., & Folstein, M. (1988). The Telephone Interview for Cognitive Status. Neuropsychiatry, Neuropsychology, and Behavioral Neurology, 1, 111–117. Braun, M., Scholz, U., Bailey, B., Perren S., Hornung, R., & Martin, M. (2009). Dementia caregiving in spousal relationships: A  dyadic perspective. Aging and Mental Health, 13, 426–436. doi:10.1080/13607860902879441 Cantor, M. H. (1975). Life space and the social support system of inner city elderly of New York. Gerontologist, 15, 23–27. Crimmins, E. M., Kim, J. K., Langa, K. M., & Weier, D. R. (2011). Assessment of cognition using surveys and neuropsychological assessment: The Health and Retirement Study and the Aging, Demographics, and Memory Study. Journals of Gerontology: Psychological and Social Sciences, 66, 162–171. doi:10.1093/ geronb/gbr048 Dassel, K. B., & Carr, D. C. (2014). Does dementia caregiving accelerate frailty? Findings from the Health and Retirement Study. The Gerontologist. Advanced online publication. doi:10.1093/ geront/gnu078 de Vugt, M. E., Jolles, J., van Osch, L., Stevens, F., Aalten, P., Lousberg, R., & Verhey, F. R. (2006). Cognitive functioning in spousal caregivers of dementia patients: Findings from the prospective MAASBED study. Age and Ageing, 35, 160–166. Fisher, G. G., Rodgers, W. L., & Weir, D. R. (2009). HRS imputation of cognitive functioning measures: 1992–2006 data description. Gallagher-Thompson, D., & Coon, D. W. (2007). Evidence-based psychological treatments for distress in family caregivers of older adults. Psychology of Aging, 22, 37–51. Garand, L., Rinaldo, D. E., Alberth, M. M., Delany, J., Beasock, S. L., Lopez, O. L., … Dew, M. A. (2014). Effects of problem solving therapy on mental health outcomes in family caregivers of persons with a new diagnosis of mild cognitive impairment or early dementia: A randomized controlled trial. American Journal of Geriatric Psychiatry, 22, 771–781. doi:10.1016/j.jagp.2013.01.007 Gaugler, J. E., Mittleman, M. S., Hepburn, K., & Newcomer, R. (2010). Clinically significant changes in burden and depression among dementia caregivers following nursing home admission. BMC Medicine, 8, 1–14. doi:10.1186/1741-7015-8-85 Gaugler, J. E., Wackerbarth, S. B., Mendiondo, M., Schmitt, F. A., & Smith, C. D. (2003). The characteristics of dementia caregiving onset. American Journal of Alzheimer's Disease and Other Dementias, 18, 97–104. Gualtieri, T. C. (2013). Husband-wife correlations in neurocognitive test performance. Psychology, 4, 771–775. doi:10.4236/ psych.2013.410109 Haley, W. E., Bergman, E. J., Roth, D. L., McVie, T., Gaugler, J. E., & Mittelman, M. S. (2008). Long-term effects of bereavement and caregiver intervention on dementia caregiver depressive symptoms. The Gerontologist, 48, 732–740. Health and Retirement Study. (2010). Public use dataset. Ann Arbor, MI: University of Michigan. Hebert, L. E., Weuve, J., Scherr, P. A., & Evans, D. A. (2013). Alzheimer disease in the United States (2010-2050) estimated using the 2010 census. Neurology, 80, 1778–1783. doi:10.1212/ WNL.0b013e31828726f5 Holley, C. K., & Mast, B. T. (2009). The impact of anticipatory grief on caregiver burden in dementia caregivers. Gerontologist, 49, 388–396. doi:10.1093/geront/gnp061

9

10

Vitaliano, P. P., Murphy, M., Young, H. M., Echeverria, D., & Borson, S. (2011). Does caring for a spouse with dementia promote cognitive decline? A hypothesis and proposed mechanisms. Journal of the American Geriatrics Society, 59, 900–908. doi:10.1111/j.1532-5415.2011.03368.x Vitaliano, P. P., Russo, J., Young, H. M., Teri, L., & Maiuro, R. D. (1991). Predictors of burden in spouse caregivers of individuals with Alzheimer's disease. Psychology and Aging, 6, 392–402. Vitaliano, P. P., Strachan, E., Dansie, E., Goldberg, J., & Buchwald, D. (2014). Does caregiving cause psychological distress? The care for familial and genetic vulnerabilities in female twins. Annals of Behavioral Medicine, 47, 198–207. doi:10.1007/ s12160-013-9538-y Vitaliano, P. P., Zhang, J., & Scanlan, J. M. (2003). Is caregiving hazardous to one's physical health? A  metaanalysis. Psychological Bulletin, 129, 946–972. doi:10.1037/0033-2909.129.6.946 Vitaliano, P. P., Zhang, J., Young, H. M., Caswell, L. W., Scanlan, J. M., & Echeverria, D. (2009). Depressed mood mediates decline in cognitive processing speed in caregivers. The Gerontologist, 49, 12–22. doi:10.1093/geront/gnp004 Welsh, K. A., Breitner, J. C.  S., & Magruder-Habib, K. M. (1993). Detection of dementia in the elderly using telephone screening of cognitive status. Neuropsychiatry, Neuropsychology, & Behavioral Neurology, 6, 103–110. Yaffe, K., Fox, P., Newcomer, R., Sands, L., Lindquist, K., Dane, K., … Covinsky, K. E. (2002). Patient and caregiver characteristics and nursing home placement in patients with dementia. Journal of the American Medical Association, 287, 2090–2097. Yang, Y., & Lee, L. C. (2010). Dynamics and heterogeneity in the process of human frailty and aging: Evidence from the U.S. older adult population. Journal of Gerontology: Social Sciences, 65B, 246–255. doi:doi:10.1093/geronb/gbp102

Downloaded from http://gerontologist.oxfordjournals.org/ at Monash University on November 21, 2015

RAND Center for the Study of Aging. (2008). RAND HRS Data, Version H. Santa Monica, CA. With Funding From the National Institute on Aging and the Social Security Administration. Roth, D. L., Fredman, L., & Haley, W. E. (2015). Informal caregiving and its impact on health: A reappraisal from populationbased studies. Gerontologist, 55, 309–319. doi:10.1093/geront. gnu177 Russo, J., Vitaliano, P. P., Brewer, D. D., Katon, W., & Becker, J. (1995). Psychiatric disorders in spouse caregivers of care recipients with Alzheimer's disease and matched controls: A  diathesis-stress model of psychopathology. Journal of Abnormal Psychology, 104, 197–204. Schulz, R., Belle, S. H., Czaja, S. J., McGinnis, K. A., Stevens, A., & Zhang, S. (2004). Long-term care placement of dementia patients and caregiver health and well-being. Journal of the American Medical Association, 292, 961–967. doi:10.1001/jama292.8.961 Seltzer, M. M., & Li, L. W. (1996). The transitions of caregiving: Subjective and objective definitions. The Gerontologist, 36, 614–626. Smith, G. E., Kokmen, E., & O'Brien, P. C. (2000). Risk factors for nursing home placement in a population-based dementia cohort. Journal of the American Geriatrics Society, 48, 519–525. Uglanova, E. A., & Staudinger, U. M. (2013). Zooming in on life events: Is hedonic adaptation sensitive to the temporal distance from the event? Social Indices Research, 111, 265–286. doi:10.1007/s11205-012-0004-1 Vitaliano, P., Echeverria, D., Shelkey, M., Zhang, J., & Scanlan, J. (2007). A cognitive psychophysiological model to predict functional decline in chronically stressed older adults. Journal of Clinical Psychology in Medical Settings, 14, 177–190. Vitaliano, P. P., Echeverria, D., Yi, J., Phillips, P. E., Young, H., & Siegler, I. C. (2005). Psychophysiological mediators of caregiver stress and differential cognitive decline. Psychology of Aging, 20, 402–411.

The Gerontologist, 2015, Vol. 00, No. 00