Insomnia In Caregivers Of Persons With Dementia: Who Is At Risk And What Can Be Done About It?

Abstract

Sleep problems are associated with mood and function disturbances in caregivers of persons with Alzheimer's disease (AD). However, the factors associated with the onset and maintenance of sleep disturbances in caregivers of persons with dementia are unknown, and little attention has been paid to treatments to improve sleep in caregivers. Here we review some of the evidence for the association between caregiver sleep problems, and caregiver and care-recipient demographic, health, and psychosocial variables. We present data from a longitudinal study that examined factors associated with self-reported sleep problems in dementia caregivers and care-recipients over a 5-year follow-up period, and describe the existing caregiver insomnia treatment literature. We conclude with recommendations for future research.

Introduction

Sleep disturbances are common among caregivers of persons with dementia. Cross-sectional studies over the past 15 years indicate that approximately two-thirds of dementia caregivers report they are having trouble sleeping. [1] But what about the other third? How do caregivers with and without sleep complaints differ from one another? How does their sleep quality change over time? What increases or decreases a caregiver's risk for developing sleep problems? Finally, how might this information guide the development of evidence-based treatments to improve caregiver sleep?

We know that sleep disturbances in caregivers can originate from a complex set of precipitating, predisposing, and perpetuating factors, including non-conducive sleep environments, poor sleep habits, cognitive hyper-arousal and rumination, care-recipient nocturnal behaviors, age-related primary sleep disorders, and other co-morbid medical or psychiatric conditions.[1] We also know that caregiver sleep, like that of other older adults, has considerable night-to-night variability,[2,3] and that caregiver reports of their own or their care-recipient's sleep quality are not always congruent with what would be expected based upon objective measures of nighttime sleep or activity.[4,5] Given the complexity of the phenomenon, caregiver insomnia research is challenging, but the personal and socioeconomic stakes are high: poor caregiver sleep has been linked to lowered immune function, elevated stress hormones, increased risk for cardiovascular disease, and risk for premature mortality.[6-9] Continuing studies into the development, maintenance, and treatment of caregiver sleep disturbances are needed.

In this paper, we will review some of the literature describing the association between caregiver sleep problems, and caregiver and care-recipient demographic, health, and psychosocial variables. We will present data from a longitudinal study that examined factors associated with self-reported sleep problems in dementia caregivers and care-recipients over a 5-year follow-up period. Finally, we will consider the caregiver sleep treatment outcome literature in light of results from these cross-sectional and longitudinal studies, and make suggestions for future research directions.

Cross-sectional correlates of caregiver sleep disturbances

A majority of caregivers of persons with dementia in the United States are older women.[10] Both increasing age and female gender are associated with a higher prevalence of sleep complaints in community-based samples.[11,12] Older caregivers are also more likely to have a variety of comorbid medical problems which, combined with the medications used to treat these problems, increase risk for development of insomnia.[13-15] Finally, primary sleep disorders such obstructive sleep apnea (OSA) and restless legs syndrome (RLS), are more common in older adults but frequently undiagnosed,[16-18] and so may play an unrecognized role in many caregiver sleep complaints.

In addition to these predisposing and precipitating demographic and medical risk factors for insomnia, a growing body of literature suggests that the unique psychosocial circumstances faced by caregivers, and their emotional and behavioral responses to these circumstances, may perpetuate caregiver sleep complaints. Caregivers are often awakened by their care-recipients at night,[19] and caregivers sleep better when they avail themselves of respite breaks away from the care-recipient.[20] However, several studies have shown that care-recipient nocturnal behavioral disturbances are not, in themselves, necessarily associated with poor caregiver sleep,[2,21] and in fact, caregiver sleep problems often continue even after care-recipients are moved out of the home or die. [22] There is also evidence that caregiver objective sleep quality is not significantly different from age-matched non-caregiver samples, although caregivers perceive their sleep to be worse.[23] Thus, what may be more important than the actual caregiving role in the development and maintenance of insomnia is caregivers’ personal interpretation or appraisal of their situation. In recent years, a number of researchers have begun to consider how modern theories of stress and coping can inform our understanding of the relationship between caregiving and caregiver health outcomes, including those important to sleep.[24]

Using structural equation modeling, Brummett et al.[25] demonstrated that being a caregiver is related to worse sleep quality, but that this association is mediated by caregiver negative affect, which is also inversely related to perceived social support. In other words, caregivers with good social support and low levels of negative affect may be less likely to develop sleep disturbances. In the Brummett et al. study, negative affect was measured based on self-reported depression, anxiety, and stress. Depression and anxiety are both well-known risk factors for insomnia,[26,27] and their prevalence in caregivers is significantly increased compared to non-caregiving adults.[28-31] In caregivers, negative affect would also commonly include nocturnal worry or rumination, grief and bereavement, caregiver vigilance or hyperalertness, and physiological arousal, all of which have also been found to relate to sleep disturbances in older adults.[7,32-34]

The interaction between caregiver sleep disturbances and negative affect potentially flows both ways, creating a negative feedback loop that can be difficult to break. Sleep fragmentation and obstructive sleep apnea in particular cause elevations in stress hormones that further exacerbate risk for development of negative health outcomes, affect, and insomnia.[35] However, it is possible that intervention strategies that target caregiver cognitions might help interrupt this cycle. Two recent studies show that caregivers with higher self-ratings of personal mastery have lower norepinephrine reactivity and PBMC beta(2)-adrenergic receptor sensitivity in response to stress than caregivers with lower mastery scores.[36,37] Although these studies did not directly measure ratings of caregiver sleep, they do provide further evidence that positive caregiver self-appraisals may reduce the likelihood of developing stress-related physiological responses that can worsen physical health, negatively impact mood and burden, and both directly and indirectly contribute to poor nightly sleep.

Identifying dyads at greatest risk: Longitudinal studies of caregiver and care-recipient sleep

The associations presented above comparing demographic, physical, and other psychosocial correlates of sleep disturbance in caregivers were derived from cross-sectional studies. However, the factors prospectively associated with the onset of sleep disturbances in caregivers of persons with dementia are largely unknown. The few longitudinal studies that have been conducted have mostly included small sample sizes, non-dementia caregivers, relatively short follow-up periods, and a limited range of comparison outcome measures. Carter measured self-reported sleep, depression, and actigraphic sleep quality in ten caregivers of cancer patients at three sampling points over ten weeks.[38] She reported that caregiver depression and sleep outcomes varied widely over the ten weeks, suggesting that accurate assessment of sleep and mood variables requires repeated measurements in this population. Fletcher et al.[39] found that baseline levels of sleep disturbance predicted evening fatigue levels, whereas trait anxiety and family support predicted morning fatigue in 60 family caregivers of cancer patients receiving radiation treatment over a 6-month follow-up period. Their analyses, which used hierarchical linear modeling to analyze repeated measures over the study period, further illustrate the inter- and intra-individual variability in trajectories of caregiver fatigue over time, but did not look at predictors of caregiver or patient nighttime sleep. Matsuda et al.[40] evaluated 103 family caregivers of persons with dementia following the care-recipient's placement in a long-term care facility. Significant decreases in the anxiety-insomnia subscale of the General Health Questionnaire were observed for caregivers whose relatives had been in residential care for at least six months (n=41); however, in this study, caregivers were only evaluated at two sampling points, no subjective or objective outcomes specific to sleep were included, and there were no analyses examining factors associated with improvement or decline over time.

To our knowledge, the largest study to date examining the factors associated with new onset of sleep disturbances in dementia caregivers and care-recipients is an unpublished study by McCurry et al.[41] that followed 164 community-dwelling individuals with AD and their family caregivers over a 5-year interval. Self-report sleep data was collected as part of a longitudinal study examining quality of life in Alzheimer's disease (AD).[42] Subjects were 164 caregivers (mean age 69.3 years, 70% female, 82% spouses of care recipient) and their family members (mean age 76.7 years, 59% male, average dementia duration 4.5 years) diagnosed with probable or possible AD.[43]

In the McCurry et al. study, caregiver/care-recipient dyads were evaluated at baseline and every six months for up to five years (mean = 2.0 years, range = 6 to 60 months). Assessments included caregiver reports about the care-recipient's sleep, cognitive and functional status, and level of behavioral disturbance. Caregivers also reported on their own nighttime sleep, mood, and burden. A Caregiver Sleep Questionnaire (CSQ) was used to rate the frequency of seven sleep problems occurring during the past month, including being awakened by the dementia patient, being kept awake at night by worry about the caregiving role, and daytime sleepiness or fatigue (Table 1). Caregivers also rated their sleep quality and whether they felt they were getting enough, too much, or too little sleep at night. CSQ scores of 12 points or higher (out of a possible 26), which was equal to the upper 25^th percentile for the sample at baseline, were considered the “sleep disturbed” range.

Table 1.

Sleep Outcome Questionnaires

Caregiver Sleep Questionnaire (CSQ)
Problem Frequency (items # 1−7)	Sleep Quality (Item #8)	Sleep Quantity (Item #9)
0 = Never	0 = Very good	0 = Enough
1 = Rarely	1 = Satisfactory	1 = Too much sleep
2 = 2−4x/month or weekly	2 = Troubled	2 = Too little sleep
3 = > once a week or daily (reverse code items #4, 6)	3 = Poor/Very bad
1. Does the care-recipient become more confused or disoriented at night?
2. Is the care-recipient's behavior at night a problem to you or others?
3. Does worrying about your caretaking role ever keep you awake?
4. When you awaken at night, can you fall back to sleep within 10−15 minutes?
5. How often do you usually nap during the day?
6. How often do you fall asleep when you want to stay awake?
7. How often do you use sleeping medication?
8. Overall, how would you rate your sleep at night?
9. Do you feel you get too much sleep, about enough sleep, or too little sleep?
Care-recipient Sleep Disturbances (PSD)
At night, does the care-recipient (check if appropriate)?:
Call out or yell?	Have "jerky legs" in bed?
Become agitated?	See or hear things?
Have interrupted breathing spells?	Snore?
Revised Memory and Behavior Problem Checklist (RMBPC), item #10 44
How often does the care-recipient wake you or other family members up at night?
0 = Never	1 = Not in the past week	2 = 1−2x in the past week
	3 = 3−6x in the past week	4 = Daily or more often

Care-recipient (AD patient) sleep disturbance (PSD) was based upon caregiver reports of two or more specific nocturnal behaviors (out of a possible six behaviors, including: calling out or yelling; becoming agitated; interrupted breathing spells; “jerky legs” in bed; seeing or hearing things, or snoring). Waking the caregiver in the middle of the night, three or more times a week, as reported on the Revised Memory and Behavior Problems Checklist (RMBPC)[44] was also considered indicative of care-recipient sleep disturbance.

Caregiver mood and burden was evaluated using the Center for Epidemiological Studies – Depression scale (CES-D) and the Screen for Caregiver Burden (SCB; objective and subjective subscales).[45] In addition, caregivers rated their quality of life using the Quality of Life-Alzheimer's Disease (QOL-AD) scale. Care-recipient cognitive and functional status was evaluated using the Mini-Mental State Examination (MMSE)[46] and the Lawton-Brody Physical and Instrumental Self-Maintenance Scales.[47] Care-recipient mood and behavior were rated using the RMBPC (memory, depression, and disruption subscales).[44] Caregiver and care-recipient demographics included age, gender, and caregiver relationship to the care-recipient (marital partner vs. not).

Cross-sectional associations with baseline caregiver sleep disturbance were evaluated with logistic regression models. Survival analyses were conducted to predict the onset of sleep disturbances in patients and caregivers over the follow-up period. For each outcome, only caregivers or care-recipients reporting no sleep disturbances at baseline (as measured by CSQ, PSD, or RMBPC criteria) were included. Patient cognitive and functional status, depression, behavioral disturbance, and caregiver depression and burden were considered as time dependent covariates, using the score at the time of the event, the preceding visit (6 months earlier), and the change from the preceding visit to the current one. Caregiver and patient age and gender, and patient marital status were also evaluated.

Study results showed that baseline caregiver sleep disturbance (CSQ 12 or higher) was associated in logistic regression modeling with higher (worse) scores on the RMBPC-memory scale (Odds ratio (95% confidence intervals) = 2.69 (1.43, 5.03)), objective caregiver burden on the SCB (OR = 2.42 (1.37, 4.27) for a 5-point change), and male care-recipients (OR = 4.46 (1.56, 12.7)). Baseline patient sleep disturbance (PSQ 2 or higher, and/or waking caregiver 3+ times/week) was associated with higher RMBPC-memory scores (Odds ratio (95% confidence intervals) = 1.53 (1.02, 2.28)), and lower caregiver quality of life ratings (OR = 0.34 (0.15, 0.79)). Older caregiver age, increased caregiver depression, and female gender were not associated with poor caregiver sleep at baseline.

Over a mean of 24 months of follow-up (range 6 − 60 months), 18% of caregivers had a new onset of sleep disturbances as measured by a total CSQ score ≥12. Current caregiver depression and previous level of objective burden were the strongest predictors of the onset of caregiver sleep problems (Table 2). Thirty-eight percent of care-recipients had a new onset of sleep problems as measured by the PSQ (2+ problems) or RMBPC sleep item (waking caregiver 3+ times/week). Current care-recipient depression on the RMBPC and higher levels of ADL impairment predicted onset of care-recipient sleep disturbances (Table 2). Caregiver and care-recipient demographics, dementia severity or duration, caregiver subjective burden, and overall level of care-recipient behavioral disturbances were not significant predictors in the survival analysis.

Table 2.

Survival analyses

Sleep Outcome	Predictor variables	HR (95% CI)
Caregiver sleep questionnaire – 12+	CES-D*(current)	1.42 (1.12, 1.80)
	SCB-Objective (previous)	2.33 (1.43, 3.78)
Patient sleep disturbances – PSD ≥ 2 and/or waking caregiver ≥ 3x/week	Lawton-Brody ADL scale (current)	1.55 (1.20, 2.00)
	RMBPC- Depression (current)	1.52 (1.09, 2.12)

HR = hazard ratio for 5 points, CES-D and SCB; one point ADL and RMBPC 95% CI = 95% confidence interval.

^*Sleep question removed.

Treatment of sleep disturbances in caregivers

A number of reviews over the past decade have examined the impact and quality of evidence-based interventions designed to improve caregiver physical and mental health outcomes.[24,48,48] However, there has been almost no research on the development of treatments to improve caregiver sleep, and in fact, almost no studies that have even included caregiver sleep outcomes, let alone targeted them for intervention.

We are aware of only two studies published to date examining the use of cognitive-behavioral insomnia treatment techniques to improve caregiver sleep. McCurry and colleagues [50] randomly assigned 36 caregivers (age 50+ years) of persons with Alzheimer's disease into either wait list control or an active treatment that consisted of standard sleep hygiene, stimulus control, and sleep compression strategies as well as education about community resources, stress management, and training in the ABCs to reduce patient disruptive behaviors. In deference to caregivers’ age and situation (which made it impossible to control the frequency and duration of nighttime awakenings that were caused by the care-recipient), no caregiver was asked to restrict their time in bed to less than 6.5 hours, and daytime naps less than 30 minutes in duration were permitted. Caregivers in active treatment showed significant differences in self-reported sleep on the Pittsburgh Sleep Quality Index (PSQI) total scores at post-treatment and 3-month follow-up, and improvements in weekly diary reports of nightly sleep percent. No significant differences between groups were observed for caregiver mood, burden, or patient behavior problems, suggesting that sleep improvements were not an artifact of improvements in care-recipient nocturnal behaviors or caregiver depression.[50]

Carter [51] subsequently used a repeated-measures experimental design to test the feasibility and effectiveness of a brief behavioral sleep intervention (CAregiver Sleep Intervention [CASI]) that included stimulus control, relaxation, cognitive therapy, and sleep hygiene elements. These standard sleep promotion recommendations were adapted to allow caregivers to set their own sleep and relaxation goals, and to implement behavioral changes at a self-selected pace. Thirty family caregivers of advanced stage cancer patients were randomized into either CASI or control; both self-report and wrist actigraphy data were collected at baseline, 3 and 5 weeks, and 2-, 3-, and 4-months post baseline. Caregivers in this study were a wide range of ages (21 − 85 years), and no data on care-recipients’ cognitive or behavioral status were reported. Improvements in both treatment groups were observed, but at 4-month follow-up, CASI subjects had significantly lower PSQI scores and greater actigraph-measured total sleep time than control subjects.

In one other relevant study, King and colleagues [52] randomized 100 older (age 50+ years) women family caregivers of persons with dementia into either a moderate-intensity exercise (brisk walking) program or a nutrition-education attention control condition, and measured stress-induced cardiovascular reactivity, sleep quality (self-rated sleep latency, duration, and sleep quality items from the PSQI), and psychological distress (stress and depression). Participants were introduced to an exercise training program over the first six weeks and then instructed to engage in a minimum of four 30- to 40-minute home-based exercise sessions per week (mostly brisk walking) for one year. Exercise produced improvements in systolic blood-pressure reactivity and PSQI subjective sleep quality scores in caregiver participants at one year. Among exercisers, improvements in sleep quality were related to reductions in perceived stress and subjective caregiver burden. Although caregivers were not recruited into this study based on level of sleep complaints, nor was sleep the specific target of intervention, results add further support to the hypothesized relationship between caregiver sleep disturbances and negative affect or appraisals, mediated by stress-induced cardiovascular reactivity.

In sum, the limited treatment literature suggests that treatment strategies that are effective for improving sleep in non-caregiving older adults (e.g., sleep restriction, stimulus control, and exercise) can also help caregivers sleep better, despite the unique lifestyle and environmental demands that caregivers face. However, the few treatments that have been developed do date have largely not included components that may be important given what is known about the correlates of sleep problems in caregivers. For example, none of the treatment studies evaluated caregiver self-efficacy, overall health (except for blood pressure in the case of the King et al 2002 study), or social support. None included cognitive training to enhance overall caregiver mastery and competence. All three included measures of caregiver depression, stress, and quality of life, but in no case did these measures significantly improve as a result of treatment, perhaps because depressed or distressed caregivers were not specifically recruited for enrollment. Only one study included any measures of physiological arousal or stress reactivity.

Furthermore, from the standpoint of the broader evidence-based sleep treatment literature, the few studies to date have other significant methodological problems. None have recruited subjects based upon research criteria for insomnia. Study sample sizes to date have been relatively small, and two of the three studies based sleep improvements solely upon caregiver self-report. Caregiver samples in the three treatment studies were also quite heterogeneous, representing a wide range of ages, relationships to care-recipients, and care-recipient diagnosis (e.g., dementia vs. advanced cancer), further limiting our interpretation of findings. Thus, although the existing literature supports the feasibility and potential efficacy of non-pharmacological treatments for insomnia in caregivers, additional research is needed to understand how caregiver sleep disturbances can best be evaluated, prevented, and reduced.

Conclusions

In this paper, we have reviewed the literature concerning sleep disturbances in dementia caregivers, including their prevalence, correlates, and treatments. We have also provided new data regarding factors associated with the onset of caregiver sleep disturbances over up to 5 years of caring for a family member with dementia at home. However, what this review has most clearly revealed is that the study of insomnia and other sleep disturbances in caregivers is still in its infancy.

Although large percentages of caregivers report that they have sleep problems, we know little about their severity, persistence, or symptom profiles. Most research to date that has looked at caregiver health and outcomes has either ignored sleep altogether, or been limited by the diagnostic and assessment strategies used to characterize it. We also know relatively little about how (or if) sleep disturbances in the context of a care giver/recipient dyad are unique. Persons with dementia and caregivers cohabitating in community settings are often negatively impacted by one another's sleep habits and disruptions. However, there is growing evidence that being awakened by one's demented or chronically ill spouse or parent does not necessarily lead to insomnia. The cross-sectional and longitudinal data presented in this paper provide evidence that caregiver depression, burden, medical morbidity, and appraisal of their situation, as well as care-recipient depression and functional impairments may be more important in the development of caregiver sleep problems than care-recipient nocturnal behavioral disturbances.

Sleep has been called the “new vital sign,” [53] because of the growing evidence of its important role to good health. Dementia caregivers, who are “the often forgotten patient” [54] because of their increased risk for medical and psychiatric morbidity, may be particularly at risk to suffer ill consequences from chronic sleep loss superimposed upon the stress of their caregiving role. Future research is needed to better identify those caregivers who currently have, or who are at risk for developing insomnia. This will require the use of standard research insomnia assessment tools that look at both nighttime sleep and the daytime consequences of sleep loss. Evidence-based interventions need to be developed that will target the physical, cognitive, and emotional factors that may be precipitating or perpetuating these sleep problems in an individualized way that is capable of addressing the unique circumstances of each caregiving dyad. Finally, in conjunction with developing individualized treatments, we need to gain a better understanding of what would be considered a clinically meaningful improvement – what will truly improve the everyday quality of life – for the caregivers with sleep disturbances that we treat.