Well-Being in Dementia: A Cross-Sectional Dyadic Study of the Impact of Multiple Dimensions of Strain on Persons Living with Dementia and their Family Care Partners

Abstract

Background and Purpose:

The impact of dementia-related stressors and strains have been examined for their potential to threaten the well-being of either the person with dementia or the family care partner, but rarely have studies considered the dyadic nature of well-being in dementia. The purpose of this study was to examine the dyadic effects of multiple dimensions of strain on the well-being of dementia care dyads.

Methods:

Using multilevel modeling to account for the inter-relatedness of individual well-being within dementia care dyads, we examined cross-sectional responses collected from 42 dyads comprised of a hospitalized patient diagnosed with a primary progressive dementia (PWD) and their family care partner (CP). Both PWDs and CPs self-reported on their own well-being using measures of quality of life (QOL-Alzheimer’s Disease scale) and depressive symptoms (Center for Epidemiological Studies Depression Scale).

Results:

In adjusted models, the PWD’s well-being (higher QOL and lower depressive symptoms) was associated with significantly less strain in the dyad’s relationship. The CP’s well-being was associated with significantly less care-related strain, and (for QOL scale) less relationship strain.

Conclusions:

Understanding the impact of dementia on the well-being of PWDs or CPs may require an assessment of both members of the dementia care dyad in order to gain a complete picture of how dementia-related stressors and strains impact individual well-being. These results underscore the need to assess and manage dementia-related strain as a multi-dimensional construct that may include strain related to the progression of the disease, strain from providing care, and strain on the dyad’s relationship quality.

Dementia is often considered to be a threat to an individual’s well-being, whether that individual is the person who receives a dementia diagnosis (Stites et al., 2017, Orgeta et al.,2015,Nagpal et al., 2014) or the family member who provides the majority of informal dementia care (Schulz et al., 2017, Cunningham et al., 2018). Multiple stressors and strains arising from the dementia context have been examined for their potential to threaten the well-being of either the person with dementia (PWD) or the family care partner (CP) (Buckley et al., 2012, Fauth et al., 2012, Judge et al., 2010). Rarely, however, have studies considered the dyadic nature of well-being by examining the PWD’s and CP’s outcomes simultaneously. In contrast to the view that well-being is an individual and independent outcome for a PWD or a CP, we take the view that the well-being of PWDs and CPs is an inter-related dyadic phenomenon (Lyons and Lee, 2018). The goal of this paper is thus to examine the well-being of dementia care dyads (PWDs and CPs).

Defining well-being.

Well-being in the context of dementia has been operationalized in various ways with no clear consensus on the definition (Cunningham et al., 2018, Tyack and Camic, 2017, Kaufmann and Engel, 2014). Several measures have been developed to capture an individual’s subjective experience of well-being in the general population, but in the dementia context researchers more often use measures of depressive symptoms and quality of life (QOL) separately or in combination with other measures to represent at least the physical and mental health aspects of well-being (Dawson et al., 2013, Stites et al., 2017). Some measures of QOL address a specific dimension, such as health-related QOL, whereas others are more global (Karimi and Brazier, 2016). When combined with a measure of depressive symptoms, the two measures together can be seen as representing physical and mental health of caregivers, which is a more comprehensive way of defining well-being than either construct alone.

Stressors and strains.

In addition to depressive symptoms and quality of life, some researchers characterize care-related strain (i.e. caregiver burden, or the subjective strain experienced as a result of providing care) as a defining component of CPs’ well-being (Arthur et al., 2017, Cunningham et al., 2018). However, according to the Stress Process Model (SPM: the model used to guide this study), care-related strain, labeled “role overload” in the SPM, is defined as a primary stressor that is predictive of well-being (Pearlin et al., 1990, Judge et al., 2010). Other stressors and strains that have been found to impact well-being include strain from the disease itself (e.g. extent of cognitive impairment) (Buckley et al., 2012, Dawson et al.,2013, Logsdon et al., 2002), relationship strain or poor relationship quality between the CP and PWD (Häusler et al., 2016, Fauth et al., 2012, Moon et al., 2016), the CP’s perceptions of physical and psychological suffering in the PWD (Schulz et al., 2017), and the amount of difficulty the CP has in finding meaning in providing care to the PWD (Quinn et al., 2010).

Dyadic context.

Frequently, dyadic studies of well-being in dementia are limited to comparisons of CPs’ proxy reports with PWDs’ self-reports of the PWD’s well-being (Buckley et al., 2012, Pfeifer et al., 2013, Orgeta et al., 2015, Nagpal et al., 2014, Logsdon et al., 2002, Conde-Sala et al., 2009). These studies have helped to establish the reliability of the PWD’s self-report, and in some cases highlighted the potential reasons behind incongruent reports between PWDs and CPs. Such studies do not inform an understanding about the inter-related nature of the well-being of both members of dementia care dyads.

Dyadic studies that simultaneously examine the well-being of PWDs and CPs are rare (Moon et al., 2016, Häusler et al., 2016, Gellert et al., 2017), and make important contributions given the opportunity to highlight how the shared context of dementia impacts the well-being of the care dyad. Häusler and colleagues (2016) studied self-reported QOL of PWDs and CPs and found that individuals’ own perceived stress was significantly correlated with QOL for both PWDs and CPs. However, Häusler et al. (2016) used different measures of QOL for PWDs and CPs in their study, which limited their ability to perform a dyadic analysis. As is the case in every dyadic study, there is likely a significant degree of shared variance that needs to be accounted for when analyzing well-being within dementia care dyads, which can only be accomplished using appropriate dyadic analytic techniques. Moon and colleagues (2016) examined QOL in dementia care dyads in a dyadic multilevel modeling analysis, but they used an incongruence model to focus on the difference between PWDs’ and CP’s self-reported QOL, rather than a dual outcomes model or actor-partner model to examine dyadic influences on the two individuals’ well-being. The findings of Moon et al. (2016) were thus related to what predicted the difference between dyad members’ well-being. Lastly, Gellert et al. (2017) examined predictors of PWDs’ and CPs’ own depressive symptoms in an actor-partner independence model. The authors found that the PWDs’ depressive symptoms were significantly predicted by CPs’ perceptions of the dyad (i.e. significant cross-partner effects for PWDs) but no crosspartner effects for CPs’ well-being, and concluded that although well-being in dementia care dyads was inter-related, distinct stressors seem to influence PWDs’ and CPs’ well-being in the dementia context (Gellert et al., 2017).

The purpose of this study was to examine the dyadic effects of multiple dimensions of strain on the well-being of dementia care dyads. The aims were:

1. To examine the impact of the shared context of dementia on both PWDs’ and CPs’ well-being simultaneously using dyadic analytic techniques.

2. To examine the dyadic influence of dementia-related stressors and strains on wellbeing for PWDs and CPs.

Method

Participants and Procedures

Recruitment for this study took place in adult inpatient acute care units in a university hospital in the Pacific Northwest of the United States. Approval was obtained from the university’s Institutional Review Board. Convenience sampling was used to enroll 42 dyads that met the eligibility criteria. Data were collected during the years 2014 and 2015.

Inclusion criteria.

PWDs were eligible if they were admitted to an acute care unit, age 65 or older, exhibited symptoms consistent with mild to moderate dementia, and self-reported a probable or current diagnosis of an irreversible progressive dementia: Alzheimer’s disease, vascular dementia, Lewy body dementia, or frontotemporal dementia. Family CPs were eligible if they were age 21 or older and designated by the PWD as the primary family CP (defined in the study as the family member most involved in care at home).

Exclusion criteria.

Dyads were ineligible if either the PWD or family CP was unable to speak English, or if the PWD had unresolved delirium or altered level of consciousness, which was assessed by the direct care nurse prior to the researcher approaching the dyad.

Screening for eligibility was completed in two steps: 1) the investigator screened patient records for eligibility criteria, and 2) the direct care nurse screened patients and family CPs for interest in the study. If the PWD and CP were interested, the researcher met with PWD and CP individually to reassess eligibility and obtain informed consent. Individual members of the dyad (PWDs and CPs) completed the interview in-person within the hospital unit without the other member of the dyad present. Response scales for each measure were provided as visual aids for participants, who answered verbally to questions read by the researcher. Verbal responses were recorded by the researcher using REDCap, an electronic data capture and storage system.

Well-Being Measures

Quality of life was measured among PWDs and CPs (each rating their own QOL) using the Quality of Life-Alzheimer’s Disease Scale (QOL-AD): a 13-item measure of quality of life designed to represent aspects of life (interpersonal, environmental, functional, physical, and psychological) that are important to QOL for adults with and without dementia (Logsdon et al., 2002, Logsdon et al., 2005). Each item is a question pertaining to the respondent’s feelings about an aspect of their life. An example item is: “How would you describe your current relationships with your friends?” Responses are rated on a scale from 1 (poor) to 4 (excellent) for a possible range of 13 to 52, higher scores indicating greater quality of life (Cronbach’s α = .77 for PWDs; Cronbach’s α = .85 for CPs).

Depressive symptoms were measured among PWDs and CPs (each rating their own symptoms) using the Center for Epidemiologic Studies-Depression Scale (CES-D) (Radloff, 1977): a 20-item measure designed to capture the existence and severity of symptomatology associated with depression as it may have occurred in the past week (e.g. poor appetite, restless sleep). Participants rated symptoms on a scale of 0 (rarely or never experienced) to 3 (experienced on 5–7 days of the past week) and items were summed for a possible range of 0 to 60, with higher scores indicating greater depressive symptoms (Cronbach’s α = .71 for PWDs; Cronbach’s α = .84 for CPs).

Dimensions of Strain

Cognitive impairment was measured in PWDs with the original 11-item Mini-Mental State Examination (MMSE) (Folstein et al., 1975). The MMSE (scale range 0 to 30, with higher scores indicating higher cognitive function) is designed for clinician assessment of orientation, working memory, language, delayed recall, and attention/comprehension. It has been used widely in research and has good reliability (test-retest r = .89) and validity (predictive and concurrent validity) among PWDs (Tombaugh and McIntyre, 1992, Mitchell, 2009, Fillenbaum et al., 1987).

Care-related strain was measured in CPs using the Role Overload scale (Pearlin et al., 1990), which assesses the extent to which the CP’s time and energy are exhausted by the demands of caring for the PWD. CPs respond to three items regarding how worn-out and overloaded their care role makes them feel using a Likert-type scale from 1 (not at all) to 4 (very much). For example, one of the items is: “you have more things than you can handle.” The items were summed for a scale range of 3 to 12 with higher scores indicating greater strain from providing care (Cronbach’s α = .75).

Relationship strain was measured among PWDs and CPs using the 5-item Dyadic Strain subscale of the Dyadic Relationship Scale (Sebern and Whitlatch, 2007). Each item on the Dyadic Strain subscale is a statement of a potential source of strain in the relationship, for which participants rated their level of agreement from 1 (strongly disagree) to 4 (strongly agree). Items are averaged for a scale range of 1 to 4, with higher scores indicating more perceived strain in the relationship (Cronbach’s α = .69 for PWDs; Cronbach’s α = .85 for CPs).

Analytic Approach

Descriptive statistics were conducted using the software program Stata, version 15 (Statacorp, 2017). Analysis of the dyadic data was conducted using multilevel modeling (MLM) and the software program HLM, version 7 (Raudenbush et al., 2011). The multivariate- outcomes model of MLM was selected for the ability to simultaneously model each dyad member’s outcome while still estimating and controlling for the degree of shared variance in the dyad (for the original description of the multivariate-outcomes model see: (Barnett et al., 1993). Full information maximum likelihood was selected for estimation of parameter values. In this study, there were less than 3% missing data on items from the outcome variables for PWDs and less than 5% missing data on items from the outcome variables for CPs. The level 1 (unconditional) models estimated the average values (fixed effects), and the variability around the averages (random effects) for both the PWD’s and CP’s ratings of their own QOL and depressive symptoms. Predictors were introduced in level 2 (conditional) models to help explain the variability around the average.

Unconditional models.

Within-dyad variation for QOL was modeled in the equation,

$$\left[{\text{QOL}}_{ij}={\mathit{\text{β}}}_{1j}\left({\text{PWD}}_{ij}\right)+{\mathit{\text{β}}}_{2j}\left({\text{CP}}_{ij}\right)+{\text{r}}_{ij}\right]$$

where QOL_ij represents the outcome score i in dyad j. PWD is an indicator variable taking on a value of 1 if the response was obtained from the PWD, or taking on a value of 0 if the response was obtained from the CP. Similarly, CP is an indicator variable taking on a value of 1 if the response was obtained from the CP, or taking on a value of 0 if the response was obtained from the PWD. The latent true scores of ratings of QOL for PWDs and CPs are represented by ${\mathit{\text{β}}}_{1j}$ and ${\mathit{\text{β}}}_{2j}$ , respectively. The within-dyad residuals, ${\text{r}}_{ij}$, are estimated separately for PWD and CP and represent the heterogeneity around the mean scores, or intercepts, for PWDs and CPs. Thus, PWDs’ average ratings of their own QOL is the sum of their latent true score (${\mathit{\text{β}}}_{1j}$) plus measurement error (${\text{r}}_{ij}$);or, CPs’ average ratings of their own QOL is the sum of their latent true score (${\mathit{\text{β}}}_{2j}$) plus measurement error (${\text{r}}_{ij}$). Depressive symptoms reported by PWDs and CPs were modeled using identical procedures.

Conditional Models.

Between-dyad variation was modeled in the equations,

$${\mathit{\text{β}}}_{\text{1}j}{\text{= γ}}_{\text{10}}{\text{+ γ}}_{\text{11}}{\text{MMSE}}_{\text{j}}{\text{+ γ}}_{\text{12}}{\text{CARE}}_{\text{j}}{\text{+ γ}}_{\text{13}}{\text{RELAT1}}_{\text{j}}{\text{+ γ}}_{\text{14}}{\text{RELAT2}}_{\text{j}}{\text{+ u}}_{\text{1j}}$$

$${\mathit{\text{β}}}_{\text{2}j}\text{=}{\text{γ}}_{\text{20}}\text{+}{\text{γ}}_{\text{21}}{\text{MMSE}}_{\text{j}}\text{+}{\text{γ}}_{\text{22}}{\text{CARE}}_{\text{j}}\text{+}{\text{γ}}_{\text{23}}{\text{RELAT1}}_{\text{j}}\text{+}{\text{γ}}_{\text{24}}{\text{RELAT2}}_{\text{j}}\text{+}{\text{u}}_{\text{2j}}$$

where the parameters for latent true scores of PWDs (${\mathit{\text{β}}}_{1j}$) and CPs (${\mathit{\text{β}}}_{2j}$) are the outcome variables, and independent variables (γ₂₁, γ₂₂, …) are introduced to explain the heterogeneity of the outcomes (depressive symptoms and QOL). Based upon the SPM and supporting literature from studies previously conducted in the community setting, the independent variables chosen for level-2 conditional models were: MMSE = the PWD’s cognitive function; CARE = carerelated strain, RELAT1 = PWDs’ perceptions of relationship strain; and RELAT2 = CPs’ perceptions of relationship strain. Tau correlations (produced from the tau matrix of variances and covariances calculated in MLM equations) were used to compare the interrelatedness of PWD-CP outcomes. Effect sizes were calculated for conditional models using $[r=\sqrt{(}{t}^2/(t^2+df))]$

Results

Patients with dementia were mean age 80±8 years, predominantly non-Hispanic white ethnicity/race (95%), had an average MMSE score of 21±4, and a slight majority (55%) were male. The most common dementia diagnosis among patients was Alzheimer’s disease (40%), followed by vascular dementia (29%). CPs were age 61±13 years, predominantly non-Hispanic white ethnicity/race (93%), mostly female (75%), and were either adult children (70%) or spouses (30%) of patients. See Table 1 for additional demographic and descriptive data.

Table 1.

Sample (n=42 dyads) demographics and measure descriptives

	PWD Mean±SD; or n (%)	CP Mean±SD; or n (%)
Age in years* (mean±SD)	79.81±7.76	61±12.95
Female	19 (45.2%)	30 (75.0%)
Education (% > high school diploma)	29 (69.0%)	36 (90.0%)
Race & Ethnicity
White (Non-Hispanic)	40 (95%)	37 (92.5%)
Black/African-American	1 (2.4%)	1 (2.5%)
Asian	1 (2.4%)	2 (5.0%)
Hispanic/Latino	1 (2.4%)	1 (2.5%)
Marital Status (% married/partnered)	17 (40.0%)	25 (63%)
Relationship to Patient
Wife	-	9 (22.5%)
Husband	-	3 (7.5%)
Adult Daughter	-	20 (50.0%)
Adult Son	-	7 (17.5%)
Daughter-in-law	-	1 (2.5%)
Dementia Type
Alzheimer’s disease	17 (40.4%)	-
Vascular	12 (28.6%)	-
Fronto-temporal	2 (4.8%)	-
Lewy Bodies	1 (2.4%)	-
Other (Mixed or Unknown Type)	10 (23.8%)	-
Cognitive Function (MMSE mean±SD, scale 0–30)	20.55±3.86	-
Care-Related Strain (mean±SD, scale 3–12)	-	7.9±2.62
Dyadic Relationship Strain (mean±SD; scale 1–4)	1.82±0.51	2.02±0.72

^*Ages 90 years and older were all recorded as 90 to protect identity. PWD = person with dementia. CP = family care partner. MMSE = Mini-mental state examination

Unconditional (Level 1) dyadic multilevel models of well-being outcomes.

To understand the impact of the shared context of dementia on both PWDs’ and CPs’ well-being (Aim 1), we first examined the unconditional dyadic multilevel models. For the QOL-AD measure (Table 2), average ratings were β_1j = 31.01±0.77 for PWDs, and β_2j = 35.15±0.88 for CPs, indicating that, on average, both members of the dyad had moderate QOL. Scores ranged from 18 to 42 for PWDs and 20 to 48 for CPs in this sample. In the unconditional model for depressive symptoms (Table 3), average ratings were β_1j = 14.55±1.03 and β_2j = 13.11±1.44 for PWDs and CPs, respectively, indicating that both members of the dyad on average exhibited some depressive symptoms, but on average fell below the CES-D cutoff of ≥ 16 for risk of clinical depression (Radloff, 1977). Scores ranged from 2 to 38 for PWDs and 0 to 36 for CPs in this sample.

Table 2:

Multilevel Models of Dyads’ Quality of Life (QOL-AD) (n = 42 dyads)

Fixed effects	β(SE)	t-ratio	Effect size (r)
PWD intercept unconditional model	31.01 (0.77)
PWD intercept conditional Model	31.01 (0.70)
Cognitive Impairment a	−0.03 (0.20)	−0.14	.02
Care-Related Strain b	0.15 (0.14)	0.53	.09
Relationship Strain - PWD a	−1.63 (1.51)	−1.08	.18
Relationship Strain - CP b	−2.29 (1.05)	−2.18*	.35
CP intercept unconditional model	35.15 (0.88)
CP intercept conditional model	35.15 (0.68)
Cognitive Impairment a	0.42 (0.19)	2.28*	.35
Care-Related Strain b	−1.05 (0.27)	−3.87***	.55
Relationship Strain - PWD a	−3.21 (1.48)	−2.16*	.34
Relationship Strain - CP b	−0.99 (1.04)	−0.95	.16
Random Effects	Variance Component	χ2
Unconditional Model PWD	15.51	121.09***
Conditional Model PWD	11.68	100.81***
Unconditional Model CP	23.71	168.06***
Conditional Model CP	11.14	99.33***

PWD = person with dementia; CP = family care partner; QOL-AD = Quality of Life Alzheimer’s Disease Scale;

^aPWD report;

^bCP report;

^*p < .05;

^**p < .01;

^***p < .001.

Table 3:

Multilevel Models of Dyads’ (n=42 dyads) Depressive Symptoms (CES-D)

Fixed effects	β(SE)	t ratio	Effect size (r)
PWD intercept unconditional model	14.55 (1.03)
PWD intercept conditional model	14.53 (0.87)
Cognitive Impairment a	0.27 (0.25)	1.08	.18
Care-Related Strain b	0.17 (0.34)	0.47	.08
Relationship Strain - PWD a	5.33 (1.87)	2.86**	.44
Relationship Strain - CP b	1.44 (1.31)	1.10	.18
CP intercept unconditional model	13.11 (1.45)
CP intercept conditional model	13.11 (1.24)
Cognitive Impairment a	−0.10 (0.34)	−0.30	.05
Care-Related Strain b	1.61 (0.49)	3.26**	.48
Relationship Strain - PWD a	3.64 (2.70)	1.25	.21
Relationship Strain - CP b	2.55 (1.89)	1.35	.22
Random Effects	Variance Component	χ2
Unconditional Model PWD	29.51	138.89***
Conditional Model PWD	17.69	98.94***
Unconditional Model CP	77.26	307.39***
Conditional Model CP	49.81	209.38***

PWD = person with dementia; CP = family care partner; CES-D = Center for Epidemiological Studies-Depression Scale;

^aPWD report;

^bCP report;

^*p < .05;

^**p < .01;

^***p < .001.

There was a significant amount of variability around the average QOL ratings for both PWDs (χ² = 121.09, p < .001) and CPs (χ² = 168.06, p < .001). Similarly, there was a significant amount of variability around the average rating of depressive symptoms for both PWDs (χ² = 138.89, p < .001) and CPs (χ² = 307.39, p < .001), indicating sufficient heterogeneity in the data to proceed to adjusted (Level 2) models. According to tau correlations, QOL reports within dyads had a low association (.13), whereas depressive symptoms had a moderate inverse association (−.39), which reinforces need to examine at level of dyad and account for the interdependence of the individuals’ outcomes.

Conditional (Level 2) dyadic multilevel models of well-being outcomes.

To understand the dyadic influence of dementia-related stressors and strains on well-being for PWDs and CPs (Aim 2), we proceeded to conditional models adjusted for other independent variables. Less CP-reported relationship strain was significantly associated with better PWD QOL, and less PWD-reported relationship strain was significantly associated with less PWD depressive symptoms (Tables 2 & 3). Less care-related strain, less PWD-reported relationship strain, and less PWD cognitive impairment were significantly associated with better CP QOL, and less care related strain was significantly associated with less CP depressive symptoms. Thus, a significant cross-partner effect was found for the dyad (Figures 1 & 2): controlling for one’s own perception of relationship strain, the perception of lower relationship strain of the other member of the dyad was significantly associated with higher QOL for both PWDs and CPs. The effect sizes for this cross-partner effect were moderate (see Table 2).

Figure 1.

Quality of life in dementia care dyads and the cross-partner effects of perceptions of relationship strain. CP = care partner; PWD = person with dementia. Cross-partner effects indicated by dashed arrows (significant associations bolded). Unstandardized beta coefficients from conditional multilevel modeling results are shown for predictors of quality of life. The tau correlation (model-based estimation in HLM) is shown between quality of life outcomes. Pearson’s correlation is shown between relationship strain perceptions. * = p < .05; ^*** = p <.001

Figure 2.

Depressive symptoms in dementia care dyads. CP = care partner; PWD = person with dementia. Significant associations bolded. Unstandardized beta coefficients from conditional multilevel modeling results are shown for predictors of depressive symptoms (CES-D). The tau correlation (model-based estimation in HLM) is shown between PWD and CP outcomes. Pearson’s correlation is shown between relationship strain perceptions. ^** = p < .01

Discussion

Dementia has the potential to impact the well-being of both PWDs and family CPs due to increased stressors and strains across many domains. Most studies take an individual approach to examining well-being in the dementia context, despite the call for a dyadic approach (Braun et al., 2009). A largely unstudied area is how the stressors and strains of an individual may influence well-being across the dementia care dyad. In this study, we found a significant crosspartner effect for both PWDs’ and CPs’ perceptions of relationship strain, which was associated with lower QOL for the opposite member of the dyad. An additional cross-partner effect was found for CPs, who rated their own QOL as significantly lower with increasing levels of cognitive impairment in the PWD. For depressive symptoms, which we also included as an outcome to represent well-being, there was a significant association between PWDs’ own perceptions of relationship strain and PWDs’ self-reported depressive symptoms, and a significant association between the CPs’ level of care-related strain and CPs’ own depressive symptoms. The results support our hypotheses surrounding the inter-related nature of well-being in dementia care dyads, and the likelihood that well-being is impacted by multiple dimensions of dementia-related strain (e.g. strain related to the progression of the disease, strain from providing care, and strain on the dyad’s relationship quality). There are several notable aspects of these findings that deserve further discussion.

Broadly speaking, family relationships are important to individuals’ well-being, and relationship strain is a stressor that has been found to directly threaten well-being in many illness contexts and across the life course (Thomas et al., 2017). Our findings add to the growing body of literature that the quality of the dyad’s relationship (an interpersonal variable) is critical to individual well-being (Fauth et al., 2012, Häusler et al., 2016). In the current study, how one’s family member (CPs were mostly adult daughters) perceived relationship strain was significantly associated with one’s own QOL, whereas how oneself perceived relationship strain was significantly associated with one’s depressive symptoms. Although we considered QOL and depressive symptoms to both contribute to the construct of well-being, these results point to a distinction between depressive symptoms (more psychologically intrinsic) and QOL, which may be more interpersonal and dependent upon how one’s partner perceives the dyad’s relationship. Relationship strain was the only strain that was significantly associated with PWDs’ well-being, controlling for care-related strain and cognitive impairment, and the effect size was moderate. Future studies targeting the well-being of PWDs may be missing an important influence if this variable is excluded.

As expected, care-related strain was significantly associated with the CP’s well-being, and indeed, the effect size for this association was larger than other strains. Some researchers have long-considered increased care-related strain (i.e. caregiver burden or role overload) to be the defining feature of CPs’ well-being; in other words, the presence of care-related strain indicates a deficit in well-being (Stuckey et al., 1996, Cunningham et al., 2018). Our study joins the alternative perspective established by the Stress Process Model (Pearlin et al., 1990), in which care-related strain is considered to be one of many antecedents that may, or may not, contribute to the loss of well-being. This distinction is important to future efforts to mitigate care- related strain and thus improve well-being in CPs. Regardless of the direction and nature of the relationship between these two variables, it is clear that care-related strain remains an urgent part of the threat that dementia imposes on CPs’ well-being, especially depressive symptoms.

From other studies one might infer that the origins of the negative impact of dementia on well-being may lie within the extent of actual cognitive deficits caused by these diseases (and the related care tasks), within the stigma associated with the diagnosis of Alzheimer’s disease and related dementias, or both (Stites et al., 2017, Werner, 2014). In the current study, the extent of cognitive impairment was not significantly associated with PWDs’ well-being, but it did have a significant negative impact upon CPs’ well-being (QOL ratings, moderate effect size). These results may reflect the lack of insight among PWDs’ surrounding the impact of dementia on their cognitive function, which is common among PWDs even in the early stages. Another explanation is that CPs’ appraisals of the illness are known to be more negative than PWDs’ appraisals, and the projection of the impending decline could be negatively impacting CPs’ well being (Logsdon et al., 2002). Education about dementia and helping CPs reframe their experience around the support that they have and the strengths of the PWD could reduce the stigma and the negative appraisals.

Limitations.

This study was limited by the small, cross-sectional sample of mostly Caucasian participants, and care partners in this study were mostly (70%) adult daughters of patients, and so the term “partner” should not be interpreted to mean spouses, who were only 30% of the sample. The results from our study appear to be in line with other larger studies of the QOL of PWDs and CPs (Logsdon et al., 2005, Moon et al., 2016) and depressive symptoms among PWDs and CPs (Gellert et al., 2017), but caution is advised in generalizing the results of this limited sample to other populations. Another limitation is the heterogeneity of concepts that fall under the same umbrella of well-being, which prevents us from making clear comparisons to previous research. Although we used a general measure of QOL that is appropriate for the dementia context (Logsdon et al., 1999), measures of QOL in other studies are often conflated with other concepts, such as perceived health status (Karimi and Brazier, 2016). As Cunningham et al. (2018) found in their review, there is no clear consensus on either the appropriate measures or the definition of well-being in dementia CPs. Lastly, the internal consistency on several measures in this study was less than desired, especially for PWDs. We believe that it may have been due to a combination of the small sample size and the small number of items per subscale on the role overload measure and the dyadic strain subscale. Thus, replication with larger, more diverse samples is an important next step to overcome these limitations and confirm our findings.

Strengths.

The major strength of this dyadic study of well-being in dementia is that it reinforces the need for a dyadic perspective of dementia care in order to gain insight into the impact of interpersonal relationships. Second, we expanded the study of well-being in dementia to the inpatient hospital setting, which is a moment in the dementia care trajectory where it may be especially critical for family and patient to come together to support each other’s well-being. Lastly, we applied theoretical work on the stress process of family carers and PWDs as a guide and learned more about the potential cross-partner effects of perceptions of relationship strain on QOL.

Conclusions.

In order to understand the impact that dementia has on the well-being of PWDs or CPs, this study suggests that an assessment of both members of the dementia care dyad may be required for a complete picture of how dementia-related stressors and strains impact each individual’s well-being. Additionally, the results of this study underscore the need to assess and manage dementia-related strain as a multi-dimensional construct that may include strain related to the progression of the disease, strain from providing care, and strain on the dyad’s relationship quality. The ultimate goal in this line of research is to strike a balance in the dyad that allows individual dyad members to sustain their own well-being without significant detriment to the other dyad member’s well-being. A promising next step for future research would be to move beyond considering individual well-being outcomes in a dyadic context, and instead consider what dyadic health and well-being looks like (Lyons and Lee, 2018) and how to achieve it among dementia care dyads.