Multiple Chronic Conditions in Spousal Caregivers of Older Adults With Functional Disability: Associations With Caregiving Difficulties and Gains

Abstract

Objectives

Multiple chronic conditions (MCCs) are common and have harmful consequences in later life. Along with managing their own health, many aging adults care for an impaired partner. Spousal caregiving may be more stressful when caregivers have MCCs, particularly those involving complex management. Yet, little is known about combinations of conditions that are most consequential for caregiving outcomes.

Method

Using a U.S. sample of 359 spousal caregivers and care recipients from the 2011 National Aging Trends Study and National Study of Caregiving, we examined three categories of MCCs based on similarity of management strategies (concordant only, discordant only, and both concordant and discordant) and their associations with caregiving difficulties and gains. We also considered gender differences.

Results

Relative to caregivers without MCCs, caregivers with discordant MCCs reported fewer gains, whereas caregivers with both concordant and discordant MCCs reported greater emotional and physical difficulties. Wives with discordant MCCs only reported a trend for greater physical difficulties. Caregivers with concordant MCCs did not report more difficulties or gains.

Discussion

Spousal caregivers with MCCs involving discordant management strategies appear to be at risk for adverse care-related outcomes and may benefit from support in maintaining their own health as well as their caregiving responsibilities.

The rising prevalence of middle-aged and older adults with multiple chronic conditions (MCCs) is a significant public health concern (Fabbri et al., 2015). Almost one fifth of adults aged 45 to 64 and up to nearly 80% of adults aged 65 and older report two or more chronic conditions such as arthritis, diabetes, and hypertension (DuGoff, Canudas-Romo, Buttorff, Leff, & Anderson, 2014; Freid, Bernstein, & Bush, 2012). Relative to having a single chronic condition, MCCs involve more complicated care coordination and medical adherence (Anderson, 2010; Fabbri et al., 2015). Having MCCs also predicts a host of adverse outcomes among aging adults, including increased functional limitations, greater healthcare utilization and costs, and elevated mortality rates (Anderson, 2010; Fabbri et al., 2015).

In addition to managing their own health conditions, aging adults commonly provide caregiving support to impaired spouses (Schulz & Eden, 2016). Although caregiving may be viewed as a natural part of marriage, spouses are a highly vulnerable group of caregivers. Compared with nonspouse caregivers (e.g., adult children), caregiving spouses are older and have less education, lower incomes, and fewer sources of support (Pinquart & Sörensen, 2011; Reinhard, Levine, & Samis, 2014). Caregiving spouses also report higher depressive symptoms than adult child caregivers (Pinquart & Sörensen, 2011), and are more likely to experience poor health that impedes their ability to provide care (Navaie-Waliser et al., 2002; Oldenkamp et al., 2016; Piette, Rosland, Silveira, Kabeto, & Langa, 2010). Caregiving may be more difficult and less rewarding when caregivers have their own MCCs. Yet, little is known about how caregivers’ MCCs may play a role in shaping positive and negative caregiving experiences. It is plausible, however, that having MCCs may be particularly challenging for caregivers faced with discordant illness management activities that require complex self-care regimens.

Utilizing a nationally representative U.S. sample, we applied a conceptual framework focused on the concordance of illness management goals and strategies (Piette & Kerr, 2006) to examine how three patterns of spousal caregivers’ MCCs (concordant only, discordant only, and both concordant and discordant) are linked to emotional and physical difficulties and gains from caregiving. We also considered whether these associations differed by caregiver gender.

Spousal Caregivers’ MCCs and Caregiving Outcomes

Consistent with stress process models of caregiving, caring for an impaired older adult is a chronic stressor that can lead to emotional and physical problems (Aneshensel, Pearlin, Mullan, Zarit, & Whitlatch, 1995; Pinquart & Sörensen, 2007; Vitaliano, Zhang, & Scanlan, 2003). Nevertheless, population-based studies have found that, relative to noncaregivers, caregivers as a whole report better self-rated health (Robison, Fortinsky, Kleppinger, Shugrue, & Porter, 2009), show less functional decline (Fredman, Doros, Ensrud, Hochberg, & Cauley, 2009), and have lower mortality rates (Roth, Fredman, & Haley, 2015). These findings are consistent with the “healthy caregiver hypothesis,” which posits that healthier people are more likely to enter and remain in the caregiver role (Fredman et al., 2009). Caregiving may also provide health benefits, such as exercise and increasing the caregiver’s motivation to stay healthy (Fredman et al., 2009). The extent to which caregivers encounter caregiving difficulties (e.g., appraising the care role as emotionally or physically challenging) depends partly on background characteristics (e.g., age, education) that reflect exposure to stressors and available resources to manage stress (Aneshensel et al., 1995). Given that many people are resilient to the challenges of providing care, it is important to identify factors that make caregiving more difficult for some caregivers.

Caregivers’ own health problems have long been studied as a characteristic that may hinder their ability to cope with caregiving stress (see Roth et al., 2015 and Vitaliano et al., 2003 for reviews). Caregivers have been viewed as “hidden patients” (Fengler & Goodrich, 1979), with many caregiving spouses managing their own health problems in addition to their care tasks (Piette et al., 2010). Indicators of caregiver health are associated with a number of care-related outcomes. Caregivers with worse self-rated health, for instance, tend to report elevated depression and their care recipient is at heightened risk for nursing home placement (Buhr, Kuchibhatla, & Clipp, 2006; Gaugler, Yu, Krichbaum, & Wyman, 2009; Pinquart & Sörensen, 2011). Likewise, caregivers with poorer functional status or a higher number of chronic health conditions appear more likely to relinquish their care role (Buhr et al., 2006; Liu, Kim, & Zarit, 2015; McCann, Hebert, Bienias, Morris, & Evans, 2004). Collectively, these studies suggest that the burden of spouses’ MCCs may make caregiving more emotionally and physically taxing. Indeed, spouses reporting poor health or a serious health problem perceive caregiving as more difficult (Navaie-Waliser et al., 2002).

Along with care-related stress, researchers have recognized the potential for caregiving gains (i.e., benefits of providing care) such as pride in coping with challenges and fulfilling a sense of reciprocity to the care recipient (e.g., Cohen, Colantonio, & Vernich, 2002; Goodman, Steiner, & Zarit, 1997; Lawton, 1991; Polenick, Leggett, & Kales, 2017; Sánchez-Izquierdo, Prieto-Ursúa, & Caperos, 2015). Caregiving gains are conceptually distinct from difficulties and may buffer the detrimental impact of care-related stressors on well-being (Lawton, 1991; Mackenzie & Greenwood, 2012). At the same time, negative caregiving experiences could be an obstacle to realizing these benefits (Pinquart & Sörensen, 2003). Caregivers’ health problems may be one aspect of caregiving that dampens positive outcomes associated with the care role.

Spousal caregivers’ MCCs may be linked to greater difficulties and fewer gains from caregiving for several reasons. First, caregivers with MCCs have higher time demands from their own and their care recipient’s health-related activities (e.g., medical appointments) than caregivers without MCCs (Jowsey, McRae, Gillespie, Banfield, & Yen, 2013). Second, caregiving diverts attention and resources from adequately caring for one’s own health problems (Tommis et al., 2009). Chronic conditions often require daily self-management activities (e.g., blood glucose monitoring for diabetes, medication for heart disease), and caregivers with MCCs may be less likely to adhere to treatment regimens than their noncaregiving counterparts. Poor self-care can also lead to weight gain or obesity, which is more prevalent among caregivers than noncaregivers and influences physiological indicators (e.g., blood pressure, lipids, inflammatory markers) that may worsen disease outcomes (Poirier et al., 2006; Vitaliano, Russo, Scanlan, & Greeno, 1996). Third, caregiving involves physical exertion that may exacerbate caregivers’ chronic conditions or their symptoms (e.g., pain), potentially complicating the management of MCCs over time (Pinquart & Sörensen, 2007; Tommis et al., 2009). Fourth, many chronic conditions (e.g., coronary heart disease, diabetes, hypertension) are associated with impaired cognition (Eggermont et al., 2012; Leritz, McGlinchey, Kellison, Rudolph, & Milberg, 2011). Care-related stress can further accelerate cognitive decline that reduces caregivers’ long-term functioning (Vitaliano, Echeverria, Shelkey, Zhang, & Scanlan, 2007), interferes with their ability to manage everyday tasks (Vitaliano, Murphy, Young, Echeverria, & Borson, 2011), and diminishes their self-monitoring capacities (Vitaliano, Ustundag, & Borson, 2017). Cognitive problems may in turn have negative implications for caregivers’ self-management of MCCs, the quality of care they provide, and overall household function and safety. Finally, physical comorbidities are linked to depression and anxiety (Bhattarai, Charlton, Rudisill, & Gulliford, 2013; Gould, O’Hara, Goldstein, & Beaudreau, 2016), which can compromise caregivers’ capacity to manage stressors.

The nature of caregivers’ MCCs may also have implications for caregiving outcomes. Piette and Kerr (2006) developed the concordant–discordant model for considering how types of MCCs affect care management. Their framework originally centered on diabetes comorbidities but has proven valuable in its applications to other disease contexts (e.g., chronic kidney disease, hyperlipidemia; Bowling et al., 2017; Lagu et al., 2008). According to this model, MCCs are less difficult to manage when they involve concordant goals and strategies. Concordant conditions reinforce medical recommendations and simplify self-care activities (Magnan et al., 2015; Sales et al., 2008). Hypertension and heart disease, for example, both emphasize cardiovascular risk reduction via addressing common risk factors (e.g., maintaining a healthy weight). By contrast, MCCs with discordant goals and strategies are more challenging to manage and are linked to poorer outcomes such as increased hospitalizations and mortality risk (Bowling et al., 2017; Lagu et al., 2008). Arthritis, for instance, requires a focus on managing pain and physical function, which can compete with efforts to minimize cardiovascular risk factors among people with comorbid hypertension. These difficulties may be compounded for individuals with both discordant and concordant conditions due to higher management burden and additional demands on limited resources. Bolstering this possibility, a study comparing three categories of diabetes comorbidities (concordant only, discordant only, and both concordant and discordant) found that having concordant and discordant conditions was linked to the highest risk of hospitalization (Petrosyan et al., 2017). Hence, spouses with discordant MCCs may be most likely to perceive greater difficulties and diminished gains from caregiving, especially when they manage both concordant and discordant conditions.

Potential Gender Differences

Caregiving and marital relationships are highly gendered (Aneshensel et al., 1995; Pinquart & Sörensen, 2006; Thomeer, Reczek, & Umberson, 2015). Men and women may therefore differ in how their own MCCs are linked to difficulties and gains in caring for their partner. Compared with men, women have more societal expectation to place their family members’ needs above their own and perceive more family-related barriers to maintaining their own health (Rosland, Heisler, Choi, Silveira, & Piette, 2010). Women are also less inclined to care for themselves when they are ill and receive less spousal and family support in managing their health problems (Rosland et al., 2010; Thomeer et al., 2015). Moreover, men generally take a managerial approach to caregiving and delegate tasks to other helpers, whereas women tend to be emotionally entrenched in the caregiver role (Thompson, 2002). In addition, women tend to provide more hours of care, help with a wider range of care activities, and experience more secondary stressors (e.g., financial problems) than do men (Pinquart & Sörensen, 2006; Swinkels, van Tilburg, Verbakel, & Broese van Groenou, 2017). Perhaps partly as a consequence, women report higher caregiver burden and depressive symptoms, lower subjective well-being, and poorer physical health (Pinquart & Sörensen, 2006; Swinkels et al., 2017). These findings raise the question of whether caregiving wives with MCCs may experience more care-related difficulties and fewer gains than their male counterparts.

The Present Study

Using a nationally representative U.S. sample of spousal caregivers and their community-dwelling partners, we examined cross-sectional associations between three different patterns of caregivers’ MCCs (concordant only, discordant only, and both concordant and discordant) and their perceptions of caregiving. We hypothesized the following:

• 1) Relative to caregivers without MCCs, caregivers with discordant MCCs report greater emotional and physical difficulties and fewer gains from caregiving.

• 2) The associations between MCCs and caregiving outcomes are particularly strong for caregivers with both concordant and discordant MCCs.

• 3) These associations will be stronger for wives than for husbands.

Methods

Participants

The sample included 359 community-dwelling adults aged 65 and older with functional disability and their spousal caregivers surveyed as part of the 2011 National Health and Aging Trends Study (NHATS) and National Study of Caregiving (NSOC). Participants were eligible for NHATS if they were Medicare enrollees, resided in the contiguous United States, and received health-related assistance in the last month with mobility, self-care activities, and/or household chores. Participants were recruited from a Medicare enrollment database using a stratified three-stage sampling design that oversampled the oldest age groups and black non-Hispanic individuals. Of the 12,411 enrollees who were contacted, 8,245 (71%) were interviewed in 2011.

NHATS participants were eligible for NSOC if they had at least one family member or nonpaid unrelated helper who provided health-related assistance. The 2,423 eligible NHATS participants had 4,935 eligible caregivers. Contact information was obtained for 3,362 (68.1%) caregivers, of whom 2,007 (59.7%) completed a 30-minute telephone interview in 2011.

A total of 422 caregiving spouses were interviewed. We were interested in caregivers who reside with their partner in the community, and so we removed 32 with partners living in residential care and two not living with their partner. Of the remaining 388 caregivers, 29 were missing data on study variables. Hence, we focused on 359 spousal caregivers (Table 1).

Table 1.

Background Characteristics and Key Variables for Caregiving Wives and Husbands

	Caregiving wives (n = 210)		Caregiving husbands (n = 149)
Characteristic	M	SD	M	SD
CG Age in years	71.51***	9.52	76.18	7.43
CR Age in years	77.90***	7.55	75.38	7.08
CG Chronic conditions	2.06**	1.33	1.66	1.26
CR Chronic conditions	3.04	1.50	3.06	1.41
CG ADL/IADL assistance	5.60***	2.45	4.49	2.38
CG Medical care activities	5.76***	3.18	4.04	2.88
CG Informal support	1.78	1.01	1.62	1.04
CG Emotional difficulties	1.70***	1.79	0.72	1.25
CG Physical difficulties	1.42***	1.78	0.52	1.10
CG Caregiving gains	3.56	0.44	3.53	0.49
	Proportions
CG Educational attainment
High school graduate	.34		.25
Some college	.19		.22
College graduate	.08		.14
Post graduate	.06		.09
CG Employed for pay	.16		.13
CG Paid help	.21		.18
CR Dementia status	.22**		.11

Note: ADL = activities of daily living; CG = caregiver; CR = care recipient; IADL = instrumental activities of daily living; SD = standard deviation. N = 359 spousal caregivers.

*Indicates a significant difference between caregiving wives and husbands at p ≤ .05.

**Indicates a significant difference between caregiving wives and husbands at p ≤ .01.

***Indicates a significant difference between caregiving wives and husbands at p ≤ .001.

On average, couples had been married for 43.05 years (standard deviation [SD] = 17.93, range = 1–72).

Measures

Caregivers’ MCCs

Caregivers reported whether they have been diagnosed with six chronic conditions: arthritis, diabetes, heart disease, hypertension, lung disease, and osteoporosis.

Six U.S. board–certified geriatricians at an academic medical center independently rated whether each of 15 possible combinations of these conditions was generally concordant or discordant regarding their management goals and strategies (overall inter-rater reliability = .96). From these ratings, two groups of conditions were categorized as concordant: (a) diabetes, heart disease, and hypertension (inter-rater reliability = .94), which are part of the same pathophysiologic risk profile, have shared treatment plans, and emphasize cardiovascular risk reduction (including both primary and secondary cardiovascular prevention; Kerr et al., 2007; Magnan et al., 2015; Piette & Kerr, 2006); and (b) arthritis and osteoporosis (inter-rater reliability = .83), which do not have the same pathophysiologic pathways but share the management goals of maintaining physical function, minimizing the risk of falls and subsequent fractures, and reducing or preventing pain (National Institutes of Health [NIH] Osteoporosis and Related Bone Diseases National Resource Center, 2016). Lung disease was not concordant with any other condition (inter-rater reliability = .93). Discordant conditions occurred when one condition was comorbid with one or more conditions that were not concordant.

We then grouped caregivers into four mutually exclusive categories (Petrosyan et al., 2017): 1 = has one or no chronic conditions; 2 = has concordant MCCs only; 3 = has discordant MCCs only; and 4 = has both concordant and discordant MCCs. We also examined the presence of 4 concordant and 11 discordant condition combinations or comorbidities (Table 2): 1 = has comorbidity, 0 = does not have comorbidity. Lastly, we considered caregivers’ total number of chronic conditions.

Table 2.

Spousal Caregivers’ Chronic Conditions by Category of MCCs

	One or no conditions (n = 145)	Concordant MCCs only (n = 38)	Discordant MCCs only (n = 80)	Concordant and discordant MCCs (n = 96)
Condition	N (%)	N (%)	N (%)	N (%)
Arthritis	28 (19.3)	9 (23.7)	68 (85.0)	89 (92.7)
Diabetes	5 (3.4)	22 (57.9)	11 (13.8)	60 (62.5)
Heart disease	6 (4.1)	12 (31.6)	2 (2.5)	33 (34.4)
Hypertension	43 (29.7)	28 (73.7)	65 (81.3)	90 (93.8)
Lung disease	4 (2.8)	0 (0.0)	17 (21.3)	32 (33.3)
Osteoporosis	0 (0.0)	9 (23.7)	8 (10.0)	38 (39.6)
Concordant MCCs
Arthritis-osteoporosis	—	9 (23.7)	—	37 (38.5)
Heart disease-diabetes	—	5 (13.2)	—	12 (12.5)
Heart disease-hypertension	—	11 (28.9)	—	29 (30.2)
Hypertension-diabetes	—	21 (55.3)	—	57 (59.4)
Discordant MCCs
Arthritis-diabetes	—	—	9 (11.3)	55 (57.3)
Arthritis-heart disease	—	—	2 (2.5)	29 (30.2)
Arthritis-hypertension	—	—	57 (71.3)	83 (86.5)
Arthritis-lung disease	—	—	11 (13.8)	26 (27.1)
Diabetes-lung disease	—	—	1 (1.3)	20 (20.8)
Diabetes-osteoporosis	—	—	1 (1.3)	14 (14.6)
Heart disease-lung disease	—	—	0 (0.0)	14 (14.6)
Heart disease-osteoporosis	—	—	0 (0.0)	12 (12.5)
Hypertension-lung disease	—		14 (17.5)	29 (30.2)
Hypertension-osteoporosis	—	—	5 (6.3)	34 (35.4)
Lung disease-osteoporosis	—	—	2 (2.5)	14 (14.6)

Note: CG = caregiver; CR = care recipient; MCCs = multiple chronic conditions. N = 359 spousal caregivers.

Caregiving Difficulties

Caregivers’ own perceptions of caregiving difficulties were assessed with two measures. Caregivers were first asked whether they experienced emotional or physical caregiving difficulties (1 = yes, 0 = no). Caregivers reporting difficulties were then asked to describe their extent from 1 (a little difficult) to 5 (very difficult). We combined these measures such that caregivers who did not report difficulties were assigned a zero and caregivers who reported difficulties received a score based on their degree of difficulties (0 = no difficulty to 5 = high difficulty). Separate scores were determined for emotional and physical difficulties.

Caregiving Gains

On a scale from 1 (not so much) to 4 (very much), caregivers reported the degree to which caregiving has: (a) made them more confident about their abilities; (b) taught them to deal with difficult situations; (c) brought them closer to their partner; and (d) given them satisfaction that their partner receives good care. Mean scores were determined (α = 0.67).

Caregiver Gender

We considered caregiver gender (1 = female, 0 = male) to examine whether the associations in this study significantly differed between wives and husbands.

Covariates

We controlled for two caregiver sociodemographic characteristics associated with caregiving stress: age and education (1 = no schooling completed to 9 = masters, professional, or doctoral degree; Aneshensel et al., 1995; Pinquart & Sörensen, 2003). We also controlled for caregivers’ support resources (Zarit & Femia, 2008), caregiving tasks (Pinquart & Sorensen, 2003; Polenick et al., 2017; Wolff, Spillman, Freedman, & Kasper 2016), and care recipients’ health conditions and dementia status (Oldenkamp et al., 2016; Spillman, Wolff, Freedman, & Kasper, 2014), all of which are linked to adverse caregiving outcomes.

We examined two types of caregiver support resources. As a measure of informal support, caregivers reported whether they have friends or relatives who: (a) talk to them about important things in their life; (b) help with daily activities; and (c) help care for the care recipient. Summed scores were calculated. Paid help was measured by asking whether caregivers had used a paid helper to assist with care tasks within the past year (1 = yes, 0 = no).

Caregivers reported on the help they gave their partner with activities of daily living (ADLs), instrumental activities of daily living (IADLs), and medical care activities. ADL assistance included help with bathing, dressing, eating, toileting, getting in/out of bed, and mobility inside and outside the house. IADL assistance involved help with laundry, shopping, preparing meals, banking, and managing money. Medical care activities included keeping track of medications, giving shots/injections, managing medical tasks (e.g., ostomy care, IVs, testing blood), helping with exercises, helping with a special diet, teeth/denture care, skin/wound care, foot care (e.g., clipping nails), ordering medication, scheduling appointments, speaking to providers, changing/adding a health insurance/drug plan, and other medical insurance matters. Summed scores were created for total ADL/IADL assistance and medical care activities.

Two indicators of care recipients’ health and functioning were measured using reports by care recipients (n = 311) or their proxies (n = 48; 46 spouses, 1 adult daughter, and 1 adult son). Care recipients’ dementia status (1 = dementia, 0 = no dementia) and number of chronic health conditions (arthritis, cancer, diabetes, heart disease, hypertension, lung disease, osteoporosis, and stroke) were determined from reports of a physician diagnosis.

Analytic Strategy

To examine associations between caregivers’ MCCs and care-related outcomes, we performed regressions using Mplus version 5.21. We adjusted for significant overdispersion related to low overall reports of emotional (α = 1.00, p < .001) and physical caregiving difficulties (α = 2.02, p < .001) by estimating negative binomial regressions for these outcomes. The model for caregiving gains was not overdispersed, and so was estimated using ordinary least squares regression. In the first step, we entered the covariates, caregiver gender, and the three dummy coded (1, 0) categories of MCCs as predictors (with caregivers reporting one or no chronic conditions as the reference group). In the second step, we entered three interaction terms for each category of MCCs (e.g., concordant MCCs only × caregiver gender) to determine whether associations varied significantly by gender. Prior to creating the interaction terms, predictor and moderator variables were effect coded (1, −1) to facilitate interpretation of the estimates. Continuous variables were mean centered. We explored the nature of significant interactions by examining simple slopes for wives and husbands.

Observations from NHATS and NSOC are weighted to provide nationally representative estimates and to adjust for differential probabilities of sample selection. We focused on caregivers as the unit of analysis. We used the caregiver response weight from NSOC and statistical procedures to account for clustering and stratification in the complex survey design (Kasper, Freedman, & Spillman, 2013).

Results

Table 1 displays caregivers’ background characteristics and scores on key variables. Relative to husbands, wives were younger, cared for older partners, were more likely to care for a partner with dementia, had more of their own chronic conditions, and reported greater emotional and physical care-related difficulties. Wives also provided more help with ADLs/IADLs and medical care activities.

Table 2 lists caregivers’ chronic health conditions by category of MCCs. The co-occurrence of diabetes and hypertension was the most common concordant comorbidity. Arthritis and hypertension was the most frequent discordant comorbidity.

Table 3 shows bivariate associations among caregiver background characteristics, categories of MCCs, and caregiving outcomes in preliminary analyses. Higher educational attainment was positively correlated with emotional caregiving difficulties (r = .11, p = .04) and negatively correlated with caregiving gains (r = −.17, p = .001). Categories of MCCs were unrelated to caregivers’ age and education.

Table 3.

Pearson Correlations Among Spousal Caregivers’ Background Characteristics, Categories of MCCs, and Caregiving Outcomes

Variable	1	2	3	4	5	6	7
1. Age
2. Educational attainment	.13*
3. Concordant MCCs only	.01	−.07
4. Discordant MCCs only	.07	.01	−.18***
5. Concordant and discordant MCCs	−.01	−.09	−.21***	−.32***
6. Emotional caregiving difficulties	−.07	.11*	−.04	−.04	.24***
7. Physical caregiving difficulties	−.02	.03	−.08	.02	.25***	.56***
8. Caregiving gains	.02	−.17***	−.04	−.10	.04	−.12*	−.01

Note: MCCs = multiple chronic conditions. N = 359 spousal caregivers.

*p ≤ .05. **p ≤ .01. ***p ≤ .001.

Associations Between Spousal Caregivers’ MCCs and Caregiving Outcomes

Table 4 presents findings from the regression models. Caregivers’ female gender, higher education, more medical care activities, and care recipients’ higher number of chronic conditions and dementia diagnosis were linked to greater emotional caregiving difficulties. Caregivers’ higher education and more medical care activitieswere associated with greater physical caregiving difficulties. Whereas caregivers’ higher education was linked to fewer caregiving gains, more informal support and more medical care activities were associated with greater gains.

Table 4.

Associations Between Spousal Caregivers’ MCCs and Caregiving Outcomes

	Emotional difficulties		Physical difficulties		Caregiving gains
Predictor	B	SE	B	SE	B	SE
Step 1
CG Gender (female)	.59**	.21	.37	.29	−.02	.06
CG Age in years	−.02	.01	−.01	.01	.000	.003
CG Educational attainment	.16***	.04	.16*	.07	−.05***	.01
CG ADL/IADL assistance	.02	.04	.05	.06	−.01	.01
CG Medical care activities	.08**	.03	.11**	.04	.02*	.01
CG Informal support	−.08	.07	−.03	.11	.11***	.03
CG Paid help	.08	.18	−.01	.20	.07	.06
CR Chronic conditions	.10*	.05	.11	.07	−.004	.02
CR Dementia status	.54***	.16	.34	.26	−.08	.07
CG Concordant MCCs	.18	.26	−.40	.46	−.12	.12
CG Discordant MCCs	−.08	.20	.19	.32	−.17**	.06
CG Concordant + discordant MCCs	.49**	.19	.68*	.32	−.06	.07
N-adjusted BIC	1031.72		905.14		11990.13
Step 2
CG Concordant MCCs × Gender	.06	.14	.22	.24	.03	.05
CG Discordant MCCs × Gender	.05	.15	.36*	.16	.01	.04
CG Concordant + discordant MCCs × gender	.11	.12	.13	.13	−.03	.03
N-adjusted BIC	1038.65		906.38		16332.60

Note. BIC = Bayesian information criterion; CG = caregiver; CR = care recipient; ADL = activities of daily living; IADL = instrumental activities of daily living; MCCs = multiple chronic conditions; SE = standard error. N = 359 spousal caregivers.

*p ≤ .05. **p ≤ .01. ***p ≤ .001.

Caregiving Difficulties

Caregivers with both concordant and discordant MCCs reported significantly greater emotional (B = 0.49, p = .01) and physical (B = 0.68, p = .04) caregiving difficulties than caregivers without MCCs. Having concordant MCCs only was unrelated to caregiving difficulties.

The association between discordant MCCs only and physical caregiving difficulties was significantly moderated by caregiver gender (B = 0.36, p = .02). Specifically, relative to having one or no chronic conditions, having discordant MCCs was marginally linked to greater physical caregiving difficulties for wives (B = 0.35, p = .06) but not husbands (B = −0.38, p = .18).

Caregiving Gains

Caregivers with discordant MCCs only reported significantly fewer caregiving gains than caregivers without MCCs (B = −0.17, p = .01). Having concordant only or both concordant and discordant MCCs was not associated with gains. None of these associations were moderated by caregiver gender.

Post Hoc Tests

To explore comorbidities that may be most consequential for spouses’ caregiving difficulties and gains, we conducted sensitivity analyses among the subsample of caregivers with MCCs (n = 214). For each outcome, we estimated models with the presence of 4 concordant and 11 discordant comorbidities as predictors (see Table 2). Concordant and discordant comorbidities were examined in separate models. Specific comorbidities may have different implications for caregiving outcomes depending on caregivers’ total number of chronic conditions; thus, models controlled for this variable along with covariates from the main analysis.

Two concordant comorbidities were independently linked to caregiving difficulties. Caregivers with osteoporosis and arthritis had significantly greater emotional caregiving difficulties than caregivers with MCCs who did not report this comorbidity (B = 0.65, p = .01). Caregivers with MCCs who reported diabetes and heart disease had significantly greater physical caregiving difficulties than their counterparts without this combination (B = 1.12, p = .01).

Additionally, two discordant comorbidities were independently associated with caregiving difficulties. Caregivers with osteoporosis and heart disease had significantly greater emotional difficulties than caregivers with MCCs who did not report this comorbidity (B = 1.15, p = .004). Caregivers with diabetes and lung disease reported significantly greater physical difficulties relative to caregivers managing MCCs without this combination (B = 0.96, p = .05). None of the concordant or discordant comorbidities were independently associated with caregiving gains.

Discussion

The combined effects of increased life expectancy and unprecedented population aging ensure that the need for family caregiving will continue to rise (Schulz & Eden, 2016). Spouses are often relied on to provide this assistance, but may have multiple health problems that make caregiving more difficult and less rewarding. This study demonstrates that spousal caregivers report greater difficulties and fewer gains from caregiving when they have MCCs that require complex management. We build on prior research by showing that the links between caregivers’ MCCs and their positive and negative perceptions of caregiving depend in part on the combinations of conditions caregivers manage.

Associations Between Spousal Caregivers’ MCCs and Caregiving Difficulties

Caregivers with both concordant and discordant MCCs reported significantly greater emotional and physical care-related difficulties than caregivers with one or no chronic conditions. Post hoc tests revealed that these associations could be partly driven by several comorbidities. With regard to emotional caregiving difficulties, concordant and discordant comorbidities of osteoporosis may be particularly impactful. Spouses with the concordant combination of osteoporosis and arthritis had greater emotional caregiving difficulties than their counterparts managing MCCs without this comorbidity. Diagnosed osteoporosis is associated with fear of falling due to the risk of debilitating spine and hip fractures, often resulting in higher depressive symptoms that could hinder one’s emotional capacity to provide care (Iaboni & Flint, 2013). These emotional challenges may be compounded for spousal caregivers who also have arthritis, which is linked to high rates of depression and anxiety among older adults (Murphy, Sacks, Brady, Hootman, & Chapman, 2012). Additionally, arthritis or related pain could heighten risk of falls and osteoporotic fractures (Arden et al., 2006; Wright, Lisse, Walitt, Eaton, Chen, 2011) that complicate osteoporosis management and further escalate caregivers’ fear of falling. The discordant combination of osteoporosis and heart disease was also associated with greater emotional caregiving difficulties, perhaps partly attributed to the link between cardiovascular disease and osteoporotic fractures (Farhat & Cauley, 2008; Gerber, Melton, Weston, & Roger, 2011; Paccou et al., 2015). Caring for an adult with late-life disability may require strenuous activities (e.g., lifting or turning the care recipient) that could further increase risk of fractures (NIH Osteoporosis and Related Bone Diseases National Resource Center, 2016), and so spousal caregivers with osteoporosis and the aforementioned comorbidities may also face substantial everyday obstacles that intensify emotional caregiving difficulties.

In terms of physical caregiving difficulties, concordant and discordant comorbidities of diabetes appear to be most consequential. Spousal caregivers with the concordant combination of diabetes and heart disease reported more physical caregiving difficulties than caregivers managing MCCs without this comorbidity. Diabetes and heart disease share associations with obesity and cognitive impairment (Eggermont et al., 2012; Leritz et al., 2011; Matheus et al., 2013; Poirier et al., 2006), which may make caregiving and self-care more physically challenging (Vitaliano et al., 1996, 2007, 2011, 2016). The serious nature and often debilitating symptoms of heart disease can also take priority over diabetes care, potentially complicating the self-management of both conditions (Kerr et al., 2007). Furthermore, caregivers with diabetes and heart disease may have problems handling the side effects (e.g., fatigue, muscle aches) of medications that are common among patients with these conditions (American Geriatrics Society, 2015) and could intensify physical caregiving difficulties. The discordant combination of diabetes and lung disease was also associated with greater physical caregiving difficulties. Diabetes is associated with impaired lung function (Gläser, Krüger, Merkel, Bramlage, & Herth, 2015) that may reduce one’s physical ability to provide care. In addition, lung diseases such as chronic obstructive pulmonary disease (COPD) are frequently treated with corticosteroids that are linked to dose–response increases in serum glucose levels, more rapid diabetes progression, and a higher risk of diabetes-related complications (Gläser et al., 2015). Taken together, these findings suggest that features of caregivers’ comorbidities beyond the similarity of management goals and strategies (e.g., disease severity, medication side effects) may be important when considering their implications for physical caregiving difficulties.

Contrasting previous work (e.g., Aneshensel et al., 1995; Pinquart & Sörensen, 2003), caregivers’ higher education was linked to more emotional and physical care-related difficulties along with fewer caregiving gains, over and above caregivers’ MCCs and contextual factors including care tasks and care recipients’ dementia status. Although more educated spouses have greater resources to manage stress, these spouses may be less resilient to caregiving challenges because they lack prior exposure to adversity (Seery, 2011). Future research should consider these findings and their implications for spousal caregivers with MCCs.

Associations Between Spousal Caregivers’ MCCs and Caregiving Gains

We found that caregivers managing only discordant MCCs perceived fewer gains from caregiving relative to caregivers without MCCs. Interestingly, none of the discordant conditions examined in post hoc tests were independently associated with gains. This suggests that the presence of discordant MCCs may be more consequential for reduced positive feelings about caregiving than having particular comorbidities.

Consistent with previous research (Pinquart & Sörensen, 2003), care-related stressors (e.g., dementia status) were generally not predictive of gains. Stressors related to caregivers’ discordant MCCs, however, could be one factor that hinders positive feelings toward caregiving. Discordant MCCs may, for example, deplete caregivers’ energy and resources such that the strains of caregiving outweigh the benefits. Positive aspects of the care role are linked to less caregiver burden and depression (Cohen et al., 2002), better relationship quality with the care recipient (Sánchez-Izquierdo et al., 2015), and lower rates of institutionalization (Mausbach et al., 2004). Hence, diminished gains among spousal caregivers with discordant MCCs may have detrimental implications for caregivers and care recipients.

Somewhat surprisingly, the association between MCCs and fewer caregiving gains was not present for caregivers with both concordant and discordant MCCs. One possibility is that caregivers with conditions that require highly complex management are more understanding of their partner’s health problems than caregivers without MCCs. These caregivers may also be more likely to derive personal growth from coping with their own poor health that extends to their views of caregiving (e.g., Leung et al., 2010). Consequently, although such caregivers may report greater emotional and physical care-related difficulties, their own experience with challenging health issues could minimize the tendency to perceive fewer gains from caregiving.

Finally, it is worth noting that performing a higher number of medical care activities was associated with greater caregiving gains, independent of contextual variables and caregivers’ MCCs. Although medical care activities were also linked to greater emotional and physical care-related difficulties, these tasks may promote positive feelings about caregiving because they involve the provision of active help in managing the care recipient’s health problems (Polenick et al., 2017; Poulin et al., 2010). Thus, medical care activities could potentially buffer some of the negative caregiving consequences for spousal caregivers with MCCs.

Gender Differences

One significant gender difference emerged such that managing discordant MCCs only was associated with a trend for more physical caregiving difficulties among wives but not husbands. In line with our prediction, this finding may be due in part to wives’ relatively greater focus on caring for family members at the expense of their own self-care (Rosland et al., 2010; Thomeer et al., 2015). When wives have chronic conditions that require discordant management strategies and complex self-care, they may be particularly likely to neglect their health in ways that tend to magnify the physical challenges of caregiving.

Contrary to the hypothesis, however, there were no gender differences in the links between caregivers’ MCCs and their emotional difficulties or gains related to caregiving. Therefore, despite the greater caregiving difficulties and higher number of care tasks reported by wives in this study and previous research (Pinquart & Sörensen, 2006; Swinkels et al., 2017), wives and husbands managing discordant MCCs appear largely equal in their vulnerability to adverse care-related experiences.

Strengths and Limitations

Strengths of this study include the nationally representative sample, data on caregivers’ chronic health conditions, and the assessment of both positive and negative caregiving outcomes.

We extend prior research by examining how combinations of MCCs are linked to caregiving difficulties and gains among spousal caregivers. As a whole, the presence of discordant conditions with the added complexity of concordant conditions shows particularly strong links to caregiving difficulties, whereas having only discordant conditions is associated with diminished caregiving gains. Consistent with the concordant–discordant model, discordant conditions seem to be especially problematic for caregivers, supporting the value of applying this framework to the caregiving context. The sample included a sizable proportion of caregiving husbands, which allowed us to evaluate gender differences. Our focus on spousal caregivers’ MCCs is timely and identifies caregiver subgroups who may benefit from support in managing their own health as well as their care responsibilities. Furthermore, contrary to previous studies that consider caregiving difficulties as categorical measures (e.g., Wolff et al., 2016), our use of mean scores permitted a more straightforward and nuanced examination of these outcomes. Studies using categorical measures often report high percentages of caregivers who experience care-related difficulties (Roth et al., 2015; Spillman et al., 2014; Wolff et al., 2016). Yet, the present findings that caregiving spouses report low levels of difficulties and high levels of gains on average indicate considerable resilience, counter to the predominant conceptualization of caregiving as highly stressful. Likewise, supporting the healthy caregiver hypothesis, many caregivers (40.4%) had one or no chronic conditions. When spouses manage multiple complex health problems of their own, however, they may be more likely to encounter negative care-related outcomes. Importantly, models controlled for sociodemographic characteristics, along with caregiving activities, care recipients’ health and functioning, and support resources. Therefore, caregivers’ discordant MCCs and the occurrence of both concordant and discordant MCCs were robust and independent correlates of adverse caregiving experiences.

We acknowledge several limitations. First, the cross-sectional analyses do not permit the determination of causal associations. The associations in this study are likely bidirectional in that more negative and less positive care experiences may contribute to the development and severity of chronic conditions. Illustrating this point, caregiver status and care-related stress have been associated with elevated biomarkers (e.g., C-reactive protein, endothelial function) of chronic illness risk (Gouin, Glaser, Malarkey, Beversdorf, & Kiecolt-Glaser, 2012; Mausbach et al., 2010; Vitaliano et al., 2003). Such links may be strengthened when caregivers have pre-existing conditions (e.g., coronary heart disease, hypertension) related to these biomarkers (Vitaliano et al., 2003), suggesting that caregiver stress may interact with MCCs to predict worse caregiving outcomes. Hence, future research should consider how the dynamic interplay of caregivers’ MCCs and physiological indicators of stress unfolds over time.

Second, we were limited by the base study measures. The NSOC assessed chronic conditions and caregiving outcomes by self-report, which may be subject to bias. Women tend to report more health problems and greater caregiving difficulties than men (Pinquart & Sörensen, 2006; Vitaliano et al., 2003), for example, and so these measures may have been overreported by wives and/or underreported by husbands. Future studies should examine objective indicators of MCCs and care-related stress. Moreover, caregivers were not selected based on their care hours. As a result, spouses varied in the amount and type of care they provided. In addition, several key variables were measured with a single item, limiting reliability. Third, there is currently no consensus on optimal categories for MCCs (Fabbri et al., 2015). Our approach may therefore not have captured all aspects of caregivers’ MCCs that are relevant to caregiving outcomes. Fourth, the findings may not generalize to specific care situations (e.g., dementia caregiving). Lastly, we did not have data on caregivers’ mental health conditions. Mental health problems are often comorbid with chronic physical illness (Bhattarai et al., 2013; Gould et al., 2016) and may amplify the experience of care-related stress and its harmful impact on medical comorbidities (Mausbach, Patterson, Rabinowitz, Grant, & Schulz, 2007). Nonetheless, this study lays the foundation for subsequent research to elucidate understanding of how caregivers’ MCCs are linked to perceptions of caregiving.

Conclusion

In sum, this study demonstrates that MCCs are prevalent in spousal caregivers of older adults with functional disability and are independently associated with less positive and more negative caregiving experiences when conditions involve discordant management strategies.

With rapid population aging, half of Americans are projected to have MCCs by 2030 (Anderson, 2010). The present findings indicate that many individuals will endure difficulties in managing their own and their partner’s health problems. These difficulties pose formidable challenges as well as considerable opportunities to determine how to best support the long-term needs of a highly vulnerable subgroup of caregivers and care recipients.

Funding

This work was supported by training grant T32 MH073553-11 from the National Institute of Mental Health (Stephen J. Bartels, Principal Investigator), the University of Michigan Program for Positive Aging (Helen C. Kales, Director), and grant KL2 TR001446 from the National Center for Advancing Translational Science (Benjamin H. Han, Principal Investigator). John D. Piette is a VA Senior Research Career Scientist. The National Health and Aging Trends Study (NHATS) and National Study of Caregiving (NSOC) are sponsored by the National Institute on Aging (grant number NIA U01AG032947) through a cooperative agreement with the Johns Hopkins Bloomberg School of Public Health.

Conflict of Interest

None reported.