“Relationships, Very Quickly, Turn to Nothing”: Loneliness, Social Isolation, and Adaptation to Changing Social Lives Among Persons Living With Dementia and Care Partners

Abstract

Background and Objectives

Persons with dementia and their care partners have known risks for loneliness and social isolation throughout the disease trajectory, yet little is described about social lives in a population heterogeneous for disease stage, syndrome type, and setting.

Research Design and Methods

We conducted a secondary analysis of qualitative interviews from multiple studies to triangulate responses from a cohort of persons with dementia (n = 24), and active (n = 33) or bereaved (n = 15) care partners diverse in setting, dementia type and stage, and life experience. Interviews explored challenges related to social lives and were analyzed thematically.

Results

Persons with dementia were on average 80 years old (range: 67–94), 38% female, and 78% diagnosed with Alzheimer’s dementia; care partners were on average 67 years old (range: 40–87) and 69% female. We identified 3 primary themes. First, dyads lost external social networks due to complex factors, including discomfort of surrounding social networks, caregiving responsibilities, and progressive cognitive deficits. Second, care partners described disruptions of meaningful dyadic relationships due to progressive cognitive and functional deficits, leading to loneliness and anticipatory grief. Third, adaptive strategies centered on care partners facilitating shared social activities and programs addressing caregiver burden. An overarching theme of disease–course accumulation of barriers to social interactions and constant adaptations was present in all themes.

Discussion and Implications

Isolation and loneliness are a shared experience and source of distress for persons with dementia and care partners. Results can inform interventions tailored to individual needs and disease stages of dyads that enhance social connectedness.

Caregivers, Cognitive impairment, Social health

Background and Objectives

Persons living with dementia and their care partners have known risks for loneliness and social isolation due to the disease’s profound impact on their shared social life (Kotwal et al., 2021; Portacolone et al., 2018; Van Orden & Heffner, 2022). Loneliness is defined as the emotional distress due to the difference between one’s desired and perceived social connections, whereas social isolation is an observable, measurable deficit in the frequency and contact with family, friends, and the community (Perissinotto et al., 2019). Persons with dementia may experience loneliness related to feeling left out of broader social network activities or changes in (or losses of) previously close relationships (Carbone et al., 2022; Kotwal et al., 2016). Similarly, care partners may experience disruptions in previously enjoyable shared social activities with care recipients, and caregiving responsibilities might reduce available time for social participation more broadly (Balouch et al., 2019; Jacobs et al., 2022). Loneliness and social isolation may harm both members of the dyad as unmet social needs can negatively affect quality of life and downstream health outcomes such as depression, anxiety, grief, sleep, cardiovascular disease, high-quality health care use, and early mortality (National Academies of Sciences, Engineering, and Medicine et al., 2020; Shaw et al., 2017; Steptoe et al., 2013).

Key knowledge gaps present barriers to effective interventions to address loneliness and social isolation among persons with dementia and care partners, although prior conceptual frameworks (e.g., Pearlin Stress Process Model; Pearlin et al., 1981) do broadly acknowledge the importance of social support and social ties to psychosocial health in dementia. First, there is less knowledge about these experiences over the trajectory of dementia. The ability to navigate and cope with social loss for both persons with dementia and care partners may be dynamic, changing as dementia progresses and as intersecting health, social, and financial stressors accumulate. This may be especially relevant in the last years of life for persons with dementia as the burden of the disease and functional impairments grows, and progresses to post-bereavement among care partners (Kotwal et al., 2021).

Second, limited work has differentiated between domains of social health, such as subjective loneliness and objective social isolation, which are related, but distinct overall markers of social health. Distinguishing between structural and functional aspects of social relationships can give insight into mechanisms underlying their connection to physical, mental, and social health (National Academies of Sciences, Engineering, and Medicine et al., 2020). Structural characteristics relate to the number and frequency of interaction with relationships (e.g., objective social connection), whereas functional characteristics relate to the behavioral or emotional resources provided by relationships (e.g., social support) and adequacy of relationships (e.g., loneliness). The lack of attention to distinct social domains is particularly relevant to older adults with moderate-to-severe dementia whose voices are rarely included in studies of loneliness and isolation. Prior work has relied on national surveys (e.g., Health and Retirement Study or National Health and Aging Trends Study) which tend not to include data about loneliness and isolation for persons with moderate-to-severe dementia due to concerns of unreliable responses from either self- or proxy-response (Wang et al., 2023). These epidemiologic studies also lack measures of dementia severity and therefore have limited ability to examine care partner experiences at different stages of dementia impairment. Consequently, prior dementia literature has focused on whether leading a socially active lifestyle can delay the onset of dementia (Donovan et al., 2017; Fratiglioni et al., 2000, 2004; Seeman et al., 2001; Shankar et al., 2013; Wang et al., 2023; Wilson et al., 2007), rather than whether social isolation or loneliness are themselves a source of distress as dementia advances. Qualitative research that foregrounds the voices of persons with dementia and care partners (dyads) is therefore essential to addressing this knowledge gap.

Third, studies have generally not considered the shared resources and social well-being among dyads, instead mainly focusing on care partners. Even as persons with dementia experience cognitive and functional decline, they may continue to contribute meaningfully to shared social well-being or social activities of dyads (Waite et al., 2022). In sum, clinicians and interventionists need data on contributors to loneliness and isolation among dyads, strategies for adapting to or managing social changes, and information about diverse experiences across different stages and types of dementia (Van Orden & Heffner, 2022). A myriad of recent interventions on loneliness and social isolation were developed in the context of the coronavirus disease 2019 (COVID-19) pandemic, which might be adapted to the unique needs of persons with dementia and care partners. Moreover, understanding how persons with dementia and care partners perceive the impact of social isolation and loneliness can inform regular screening and identification as changes occur along the disease course, as social needs are typically under-recognized.

We, therefore, examined the perspectives of persons with dementia and care partners on changes to their social lives, focusing on experiences with loneliness and social isolation. We draw on the interactive biopsychosocial model (IBM) to provide a framework for considering social well-being as a shared resource between dyads (Waite et al., 2021). In the IBM model, each member of a dyad contributes psychocognitive, social, and biophysical resources to one’s own health needs as well as the shared health and social well-being of the dyad. Although individual resources can help each member of the dyad compensate for declines in the other, low resources in each domain can take a toll on both or individual members of the dyad. To ensure representation of a diverse group of persons with dementia and care partners, we analyzed interview data from two studies focused on quality of life from the perspective of persons with dementia and care partners.

Research Design and Methods

Overview

We analyzed semistructured interviews with persons with dementia and care partners from two of our studies to examine social change, isolation, and experiences of loneliness. First, the Dementia Palliative Care (DPC) study examined unmet palliative care needs and opportunities for intervention within a specialty dementia clinic affiliated with an Alzheimer’s Disease Research Center (Harrison, 2022; Shafir et al., 2022). Second, the Music and Dementia Caregiving (MDC) study examined the role of music and caregiving relationships in everyday life for persons with dementia and their care partners in the home setting (Allison et al., 2022). Both studies were approved by the University of California, San Francisco, CA, institutional review board. Key aspects of each study are available in prior publications in accordance with the Consolidated Criteria for Reporting Qualitative Studies (COREQ; Allison et al., 2022; Harrison, 2022; Shafir et al., 2022). Combining these two data sets for this secondary analysis allowed us to triangulate across multiple recruitment settings in the San Francisco Bay Area, by participant types, and by study designs and to validate findings across a more diverse sample.

DPC Study Setting, Participant Recruitment, and Data Collection

Participants were recruited from a specialty dementia center located within a quaternary medical center that offers diagnostic and longitudinal care for persons with cognitive and behavioral symptoms and their families. Participants included (1) person with dementia with mild-to-moderate disease who could self-report; (2) current care partners of community-dwelling person with dementia with mild-to-moderate disease; or (3) bereaved care partners of a person with dementia who had died 3–12 months earlier who could reflect on the entire disease journey. We identified potential participants through chart review and approached clinicians for referral and approval. Participants needed to speak English, be able to consent in writing by person or proxy, and to have been seen at least once at the specialty dementia center.

An interdisciplinary team of social scientists and physicians in behavioral neurology, geriatrics, and palliative medicine developed and piloted the interview framework and conducted preliminary analyses. Between November 2018 and September 2019, a PhD sociologist (SBG) conducted semistructured interviews in person or by phone with 9 persons with dementia, 16 current care partners, and 15 bereaved care partners. Questions focused on key experiences along the disease trajectory of persons with dementia, challenges and sources of distress, and supports that did or could have helped; example questions and probes related to social needs are provided in Supplementary Material Section 1. Interviews were transcribed verbatim, deidentified, and uploaded into ATLAS.ti qualitative data analysis and research software. Four authors (K.L. Harrison, M. Halim, C.S. Ritchie, S.B. Garrett) read all transcripts and developed codes to identify challenges or supports and related to relationship or social changes, and M. Halim coded all transcripts. Data coded with “social change” were exported for secondary analysis.

MDC Study Setting, Participant Recruitment, and Data Collection

For the MDC study, English-speaking participants were recruited through two University- and one VA-based geriatrics clinical practices in collaboration with their primary care teams. MDC was a dyadic study, where persons with dementia were recruited first and then a primary live-in care partner was identified. For persons with dementia who had professional caregivers, family members were also invited to participate. Participants included (1) persons with dementia with diagnosis of any type of dementia at any stage who had live-in care partners who were willing to participate, and (2) family members who were not the primary care partner but who wanted to participate in interviews. Participants needed to live in the San Francisco Bay area, to speak English, and to be able to consent in writing by person or proxy.

Between November 2018 and January 2020, a PhD ethnomusicologist (JG), a PhD ethnomusicologist/MD geriatrician (TA) and a research staff member with a Master’s in Psychology (AR) conducted in-home semistructured interviews and participant observation with 21 dyads. For the present study, we included 13 dyads that had a family care partner, and the 4 professional caregiving dyads in which a family care partner participated in the study. We excluded four dyads where no family members were available, and the caregiver was a hired professional because their social experiences were fundamentally different. Questions focused on the lived experiences of everyday dementia caregiving and the role of music and other meaningful activities as part of everyday life. We took a life-course approach, enabling participants to reflect on the disease trajectory of persons with dementia, challenges and sources of distress, and sources of support. We used ethnographic interviewing, which relies on open-ended questions and dialogue to create rapport and allow participants to control topics discussed (Schensul & LeCompte, 2012); example questions and probes related to social needs are provided in Supplementary Material Section 1. Fieldnotes were written following each visit and audio-recorded interviews were transcribed verbatim, deidentified, and uploaded into ATLAS.ti qualitative data analysis and research software. Two researchers (TA, JG) read all transcripts and developed a comprehensive codebook in discussion with a larger interprofessional team that had expertise in the humanities and social sciences, cognitive neuroscience, epidemiology/dyadic research, geriatrics, and palliative medicine. TA coded all transcripts using ATLAS.ti.

Unified Data Analysis

To harmonize the data, we adopted a framework method for analysis that included immersion in the data, coding to label units of meaning, examining patterns, diagramming, and interpreting data (Gale et al., 2013).

A unified secondary qualitative analysis was performed by AK, a geriatrician and palliative care physician with expertise in social isolation and loneliness and training in qualitative research methods who was not involved in either parent study. Though AK primarily worked with coded and exported data subsets, he first immersed himself in the data by reading transcripts to get the context of stories and check for key data relevant to the secondary analysis not otherwise captured in the coded exports. He met with KH and TA to understand nuances of study design, data collection, and earlier-stage analysis. Next, AK worked within the subset of data coded with social change and exported to Excel, using an inductive, iterative process to generate initial codes. Members of the original research team (KH, MH, TA) reviewed and discussed new codes and code applications until consensus was achieved. For example, data that AK coded with “Cognition impacting socializing - loss of verbal skills,” KH had coded with “person living with dementia unable to interact/engage socially”; AK had also developed a code for “Inappropriate social behaviors” and combined these categories into a code for “Impaired social skills.” The team agreed on a code name of “Barriers to socializing from cognitive loss.” AK then aggregated codes into major themes (e.g., the aforementioned example became part of a theme of “Loss of external social networks”). AK also generated tables to summarize data relevant to each theme and a figure diagramming findings within the IBM. Draft tables and figures were iteratively discussed with members of the original research investigators (DPC: KH, MH, CSR, SG, TA; MDC: TA, KH) and additional team members with relevant expertise (CP, AS) to identify and resolve discrepancies until consensus was achieved.

Results

Demographics for each study sample is displayed in Table 1. Twenty-six persons with dementia participated (age 67–94). Forty-eight care partners participated, 33 of whom were actively providing caregiving, and 15 of whom were bereaved. In the DPC study, participants included three spousal dementia caregiving dyads; the remainder were with unrelated persons with dementia and care partners. For the MDC study, all participants were family dementia–caregiving dyads.

Table 1.

Demographics of Interview Participants, Including Persons Living With Dementia, Current Care Partners of Persons Living With Dementia, and Former Care Partners of Decedents Who Died With Severe or Advanced Disease

	Dementia Palliative Care Study			Music Dementia Caregiving Study
	Persons living with dementia (n = 9)	Current care partners (n = 16)	Former care partners (n = 15)	Persons living with dementia (n = 17)	Current care partners (n = 17) missing age for four adult children
	n (%)	n (%)	n (%)
Self-reported demographics
Age in years (mean [range])	75.3 [67–86]	65.8 [40–87]	68.9 [45–82]	82.8 [71–94]	68.3 [47–83]
Sex
Female	3 (33%)	12 (75%)	9 (60%)	7 (41%)	12 (71%)
Male	6 (67%)	4 (25%)	6 (40%)	10 (59%)	5 (29%)
Marital status
Married/partnered/living together	6 (67%)	13 (81%)	2 (13%)	8 (47%)	12 (71%)
Widowed	—	—	11 (73%)	3 (18%)	0 (%)
Divorced/separated/never married	3 (33%)	3 (19%)	2 (13%)	3 (18%)	0 (%)
Missing	—	—	—	3 (18%)	4 (24%)
Race and ethnicitya (multiple possible)
White	7 (78%)	9 (56%)	14 (93%)	9 (53%)	7 (41%)
Black/African American	—	2 (13%)	1 (7%)	4 (24%)	4 (24%)
Asian	1 (11%)	3 (19%)	—	4 (24%)	4 (24%)
Another race or ethnicity [text available] or missing	1 (11%)	2 (13%)	—	0 (0%)	2 (11%)
Education (highest completed)
High school graduate, some college, or vocational	1 (11%)	4 (25%)	2 (13%)	8 (47%)	6 (35%)
College graduate	1 (11%)	4 (25%)	6 (40%)	1 (0.6%)	4 (24%)
Masters/PhD/professional degree	7 (78%)	8 (50%)	7 (47%)	7 (41%)	3 (18%)
Missing	—	—	—	—	4 (24%)
Clinical data abstracted from chart for recruitment
mean MMSE or MoCA (before recruitment) [range]	23.6 [15–29]	17.4 [15–29]	—	FAST stage Mild: 4 (24%), Moderate: 10 (59%), Severe/Hospice-eligible: 3 (18%)
Dementia syndromes	Clinical diagnosis	Associated dementia diagnosis	Associated dementia diagnosis	Self-report
Alzheimer dementia	5 (56%)	9 (56%)	5 (33%)	6 (35%)	0 (0%)
Lewy body	—	2 (13%)	1 (7%)	3 (42%)	0 (0%)
Behavioral variant frontotemporal dementia	1 (11%)	1 (6%)	1 (7%)	0 (0%)	0 (0%)
Other syndromesb	1 (11%)	2 (13%)	2 (13%)	1 (0.6%)	1 (0.6%)
Mixed syndromesc	2 (22%)	2 (13%)	6 (40%)	—	—

Notes: FAST = Functional Assessment Staging Tool; MMSE = Mini-Mental Status Exam; MoCA = Montreal Cognitive Assessment.

^aRace and ethnicity were collected to report per funder requirements and categorized based on NIH reporting categories.

^bOther syndromes: Person Living with Dementia: 1 nfvPPA/PSP (nonfluent variant primary progressive aphasia/progressive supranuclear palsy); care partners: 2 lvPPA (logopenic variant primary progressive aphasia), 2 PSP (progressive supranuclear palsy), 2 CBS (corticobasal syndrome—1 mixed), 2 PCA (posterior cortical atrophy—1 mixed).

^cMixed syndromes or copathologies: Person Living with Dementia: 2 AD/VD (Alzheimer’s disease/vascular dementia); care partners: 1 CBS/PCA (corticobasal syndrome/posterior cortical atrophy), 3 AD/LBD (Alzheimer’s Disease/Lewy Body Disease), 1 AD/PCA (Alzheimer’s Disease/posterior cortical atrophy), 1 lvPPA/AD (logopenic variant primary progressive aphasia/Alzheimer’s Disease, 1 early onset Parkinson’s with late onset Parkinson’s dementia.

We analytically identified three primary themes: (1) loss of external social networks among persons with dementia and care partners, (2) losses within the dyadic relationship, and (3) social adaptations made by individuals and dyads in response to disease or caregiving. An overarching theme of constant change and disruption to social lives across dementia subtype and disease stage, is highlighted within each theme. We present an updated IBM Conceptual Framework to reflect our findings (Figure 1). Quotes are cited by the study (DPC or MDC), participant number, participant type (person with dementia, care partner, or dyad), gender, age, dementia type (Alzheimer’s disease [AD] or non-AD), and dementia severity (DPC study: MMSE, MoCA, or Bereaved Care Partner; MDC study: FAST Stage). For assertations supported by quotes presented in tables, we cite the quote location in the table as follows: Table 2, Quote A (Quote 2A).

Table 2.

Loss of External Social Network of Person Living With Dementia and Caregiver

Subthemes	Quotes
Avoidance, social exclusion, and withdrawal	A. Fear/Avoidance by Surrounding Social Network: My dad, especially when my mom was gone, was very scared and emotional and labile, and easy to trigger, and so a couple friends came by and got sort of scared off … So very quickly, people didn’t come. So he was pretty lonely, except for the caregiver and me. (DPC c08, bereaved care partner, female, 45, AD)
	B. Social Exclusion of Dyad: One part that was difficult for me in the early years, that all of our friends were keeping going and doing all the things we used to do together and we were not … They weren’t purposefully doing that to exclude us, but the disease excluded us from being a part of a friendship group that continued. (DPC c20, bereaved care partner, female 72, non-AD) C. Social Exclusion of Person Living with Dementia: So she did very well [at the facility], but her dementia kept going, and something unfortunate happened … She had a group of women that she had every meal with and they started shunning her … Because her dementia got to the point where she couldn’t relate to them, and they didn’t want her sitting at their table. (DPC c11, bereaved care partner, female, 78, AD)
	D. Person Living with Dementia Withdrawal: She has really withdrawn socially from her friends and from everything. … because she said people will talk about you, and that’s her concern … She doesn’t go to church like she used to … She’s just really withdrawn. (DPC c23, care partner, female, 50, AD, MMSE 16) E. Caregiver Withdrawal: My own social life just completely deteriorated … There’s the disease and everything that you do, and then [[Spouse]] passes away, and while I know it would be good for me to be more social, I don’t have the desire to be at this point. (DPC c16, bereaved care partner, male, 61, non-AD) F. Caregiver Responsibilities: I got a dog because I don’t leave anymore. That’s a problem. Researcher 1: How come you don’t leave anymore? Caregiver: Because I have to-- because I’m the point person. (MDC, c13, bereaved care partner, female, AD)
Barriers to socializing from cognitive loss	G. Memory: “I mean, he just … he’s losing it. He’s losing people … so what am I going to do next? I’m going to do a photo album. For him. But I want him <laughs> to be able to recognize his children and his grandchildren, but even when I say, you know, ‘[[family member #5]] coming, [[family member #5]] and [[family member #6]] are coming,’ … I have to do a little bit of explanation.” (DPC c19, care partner, female, 72, AD, MMSE 19)
	H. Verbal Skills: “I try to speak less with telephone messages, and I try to avoid social gathering. My best friend before since first grade lives in [[Canadian city]]. She and I converse over an hour every Sunday afternoon, and I stopped that because I get tired…” (DPC, p10, person living with dementia, female, 77, non-AD, MMSE 28)
	I. Difficulty with Groups: “Mother’s Day was a complete disaster because he can’t be around a bunch of people. The stimulation’s too much, and he gets confused really easy. … So he got really aggravated at Mother’s Day at my sister’s, so we had to leave.” DPC c22, care partner, female, 57, non-AD)
	J. Inappropriate Social Behaviors: Well, the social issues are what are the worst, because [[Husband]], he says inappropriate things in front of people, … especially when it comes to young women, and so we don’t do a lot of social things. (DPC c29, care partner, female, 64, non-AD, MMSE 24)
	K. Safety/Independence: “Every social thing she does I have to go or she’s not comfortable … Events, community, she doesn’t participate. She used to be a big concertgoer even when she was a hundred years old. She doesn’t say, ‘I’m scared,’ but she won’t go, and if it’s more than five minutes from her house she can’t walk home, and so she’s not going.” (DPC c24, care partner, female, 59, AD, MMSE 23) L. Inability to Drive: “Not being able to drive is a biggie. I’ve gone to two Alzheimer’s group meetings. That comes out real strong, that not being able to drive … it’s a big deal for me, especially where we live. I’m not going anywhere, almost nowhere. (DPC, p04, person living with dementia, male, 73, AD, MMSE 27) M. Functional Impairment: “I’d like to get out of the house, but it’s too much work, you know. I’ve got to have my son to get me ready and have somebody to help my son to get me down and to get me up.” (MDC, p08, person living with dementia, male, 82, FAST Stage 5 (mild))
	N. Apathy: “He’s isolating himself. Very much so. I have to make him go out with me. I mean, I even try and get him out to occasionally go to see a movie or something. He doesn’t want to go, and he, honestly, he sits on a couch and watches TV all day. …. we have a couple that’s a really old friend of ours and they know what’s going on, so occasionally we’ll go out to dinner with them, and they’re so nice and kind about everything, but it’s hard to even get [[Spouse]] to have dinner with them.” (DPC c10, care partner, female, 71, non-AD, MMSE 23)

Notes: AD = Alzheimer’s dementia; DPC = Dementia Palliative Care Study; MDC = Music and Dementia Caregiving Study; MMSE = Mini-Mental Status Exam; MoCA = Montreal Cognitive Assessment; non-AD = non-Alzheimer’s dementia (includes dementia conditions listed in Table 1).

Figure 1.

Study themes in the interactive biopsychosocial model. The Interactive Biopsychosocial Model conceptualizes health and social well-being as produced in intimate, family, or social relationships. Each individual in the dyad is represented by a circle with endowments in three domains of resources, or capital: (1) biophysical (e.g., physical function, sensory function, absence of disease), (2) psychocognitive (emotional and cognitive function), and (3) social (e.g., social ties, social supports, participation in social activities, satisfaction with social relationships). In this model, the combined resources produce the “Social Capital of the Dyad.” Study findings were used to adapt the model to the dyadic experience of dementia. Dementia negatively affects the social capital of the dyad through the loss of shared external social networks and disruptions of the intimate Persons Living with Dementia–Care Partner relationship. Coping strategies involve care partner investment in facilitating social interaction with external networks, shared social activities, or paid/external support services. An overarching theme is change experienced at different stages of dementia, with accumulating or intersecting barriers to social lives, and need for continuous adaptation and creativity.

Theme 1: Loss of External Social Networks (Table 2)

Persons with dementia and care partners experienced a gradual loss of social connections as the disease progressed which affected overall well-being and the ability to mobilize supportive resources. A major contributor to this loss was avoidance and social exclusion from surrounding social networks and withdrawal from social activities. Care partners noticed the discomfort and fear that previously close family and friends experienced around persons with dementia, particularly for those whose symptoms included behavioral changes and/or labile mood (Table 2, Quote A [Quote 2A]). As one care partner noted:

The hardest part was seeing people not come back to see him … relationships, very quickly, turn to nothing … No one wanted to go in the door, you know? The couple people that went in, and it was very few, like within two to three visits were yelled out. And they just stopped coming. (DPC c08, Bereaved Care partner, female, 45, AD)

Participants described intentional and unintentional social exclusion from previously enjoyable social activities as the disease progressed (Quotes 2B and 2C). In addition to the actions of surrounding social networks, persons with dementia (Quote 2D) and care partners expressed shame and guilt about the disease symptoms. One person with dementia noted:

It’s embarrassing. You know, I think of it —a—I think of people that have Alzheimer’s as just slobbering, babbling idiots, and social, same thing, why would I want to go out and look like a slobbering, babbling idiot. (DPC p04, person with dementia, male, 73, AD, MMSE 27 points)

Care partners’ guilt centered on burdening friends and family with their presence or asking for help, and consequently actively avoided others (Quote 2E). Mounting caregiving responsibilities with greater disease severity further interfered with time or desire to socialize, work activities, and social identities (Quote 2F).

Progressive cognitive decline presented barriers to social engagement for both persons with dementia and care partners. For both members of the dyad, memory deficits in persons with dementia, for example, forgetting the names of children and grandchildren, made interactions with family and friends awkward (Quote 2G). Disrupted verbal skills, especially for semantic-variant subtypes made persons with dementia appear socially withdrawn (Quote 2H), presented barriers to following conversation in group settings (Quote 2I), and were associated with disruptive or socially inappropriate behaviors (Quote 2J). One care partner noted:

The biggest fear she has … is she feels she is very, very unsocial. She doesn’t like crowds. She doesn’t like easy conversation. (DPC c18, bereaved care partner, female, 87, AD)

Safety concerns due to functional impairment, driving, or declining executive function and judgment made previously enjoyable social activities out of reach for persons with dementia who could not independently leave the house (Quotes 2K, 2L, and 2M). Moreover, apathy presented barriers to person living with dementia initiating social activities (Quote 2N).

Theme 2: Losses Within the Dyadic Relationships (Table 3)

Over the disease trajectory, care partners described multifaceted challenges stemming from changes in their prior relationship roles with their care recipient. Care partners described subtle changes in relationship roles early in the disease course; for example, persons with dementia could no longer be the main social organizer (Table 3, Quote A [3A]). Over time, this progressed to the care partner being the sole decision maker (for social and all other activities) (Quote 3B), a lonely role. Behavioral changes in persons with dementia (Quote 3C), particularly with frontotemporal dementia but present in other subtypes too, presented barriers to spending meaningful time together. Changes contributed to care partners’ guilt around engaging in independent social activities:

Table 3.

Losses Within the Dyadic Relationship

Subthemes	Quotes
Changing relationship roles	A. Person with Dementia used to be main social organizer: [[PtName]] was in charge of our social life. She was our social director … [[PtName]] made … the plans … and then, we’d go do it … And now, I’m in charge of—if there’s some music or a play or something for the day, I have to be the one to put that together. (DPC c30, care partner, male, 67, non-AD)
	B. Loneliness of being sole decision maker: There’s just so…there’s so much that happens in the course of caring for somebody … It’s just extremely lonely because, especially when you’re making decisions for somebody you love … it’s very emotionally charged and so gee, I’m glad I had the strength to do it. (DPC c16, bereaved care partner, male, 61, non-AD)
	C. Behavioral changes impacting relationship: He could give less back. You know, going over there and, like < chokes up > sitting and not feeling like you’re doing much or helping the situation or helping him … I wondered, like, sometimes he would yell at me and, you know, yell me out of the house, and it’s like I knew he didn’t mean it and I knew it was the disease, but, you know, “Maybe I should just not go, because I clearly agitate him,” and yet if I don’t go, nobody goes. (DPC c08, bereaved care partner, female, 45, AD)
	D. Separate living situations: It was harder for me than it was for [[Spouse]]. [[Spouse]] adjusted to [[facility]] over the course of a few months and, but my husband leaving home was a sad, one of the saddest, if not the saddest day of my life. And then coping with that, with him still being alive but not being able to be at home, it was very lonely … it was much more difficult for me, being lonely, and I was exhausted from caring for [[Spouse]] at home. ((DPC c16, bereaved care partner, male, 61, non-AD)
Loneliness due to Grief and Loss of intimate relationship	E. Loss of companionship: I think just the constancy of it. It’s a never-ending task. You wake up every day not knowing what it’s going to be like, whether it’s going to be good day or a hard day … For me … the emotional stress is very debilitating because as I have always said you’re taking care of someone who’s the love of your life so you’re losing that core relationship by becoming a caregiver. But you try to cling to it and they can’t respond to it and it’s just this back and forth struggle to make sense of this relationship that you’re now basically a nurse for. And you know that your partner doesn’t really comprehend everything. They can’t articulate … he would never say, “How are you doing? How are you feeling?” He wasn’t capable of it. And I knew that intellectually but in my heart it was just really hard. (DPC c04, bereaved care partner, male, 67, non-AD) F. Loss of companionship: Interviewer: What are some of your biggest concerns, thinking of the future? Person Living with Dementia: Probably concerned about losing the relationship with [Wife]. Interviewer: Do you feel like saying a little more about that? Person Living with Dementia: Well, when you’ve been together for < laughs > more years than you can count anymore—we’ve had what I consider to be a really great marriage and relationship, and I want that to continue just as long as possible, but I’m afraid that that’s going to be shortened from what I would like it to be. (DPC, p01, person living with dementia, male, 67, AD, MMSE 24) G. Loss of companionship: Well, his dementia. It just prevents him from being the person he used to be. He’s a total stranger to me now. Honestly, it’s not him anymore. It really isn’t. It really hurts me to talk about this. I’m trying not to cry. (MDC c01, care partner, female, 65, AD, FAST Stage 6 (moderate))
	H. Anticipatory and post-death grief: Kind of constantly living with knowing where it was going, you just didn’t know when. But so, you know, I think I was probably going through an extended slow motion grieving process at the time and at some point … (DPC c13, bereaved care partner, female, 73, AD) I. Anticipatory grief: Out-of-state daughter explained that she only visits twice a year because “in the past few years and especially now it’s hard to see the … it was hard to see her. It was hard to see their decline. So it was very difficult. It was very difficult for me to see them the way that they are, knowing the way that they were.” (MDC, c03, daughter, female, age unknown, AD, FAST Stage 6 (moderate)).

Notes: AD = Alzheimer’s dementia; DPC = Dementia Palliative Care Study; MDC = Music and Dementia Caregiving Study; MMSE = Mini-Mental Status Exam; MoCA = Montreal Cognitive Assessment; non-AD = non-Alzheimer’s dementia (includes dementia conditions listed in Table 1).

I mean, I try to be here when I’m not working. But, I feel guilty about going out by myself. (DPC c25, care partner, female, 74)

In a subset of participants, disease progression and caregiver responsibilities led to the care partner’s physical separation from the persons with dementia due to moving to a care facility, an event accompanied with relief from physical burdens of caregiving, but profound social loss (Quote 3D).

Taken together, changing relationship roles generated substantial loneliness as care partners and persons with dementia cope with the loss of companionship from their spouse (Quotes 3E, 3F, and 3G). Spousal caregivers described the loss of intimacy of their “core relationship by becoming a caregiver.” One spousal care partner noted, for example:

I also feel very emotionally lonely because [[Spouse]] can’t talk to me. There’s no intellectual discussion at all, and no meaningful discussion about anything… I feel almost like I’m not married. (DPC c10, care partner, female, 71, non-AD, MMSE 23 points)

This “extended, slow-motion grieving process” was emotionally taxing for care partners and led to social withdrawal (Quotes 3H and 3I).

Theme 3: Individual and Dyadic Adaptive Strategies (Table 4)

Care partners outlined several adaptive strategies to maintain the shared social well-being for themselves and persons with dementia. Care partners facilitated social interaction by encouraging members of their social network to visit (Table 4, Quote A [4A]), creating comfortable social spaces for persons with dementia (i.e., smaller groups, familiar faces, and environments; Quote 4B, and actively addressing social discomfort [Quotes 4A and 4C]). For example:

Table 4.

Individual and Dyadic Coping Strategies Among Care Partners and Persons Living With Dementia

Subthemes	Quotes
Reliance on primary caregiver to facilitate social interaction	A. Care partner helping to overcome external network avoidance: I think it’s such a challenging part of cognitive diseases because, no judgment, but some people just don’t feel comfortable for whatever reason to even show up … I really had to encourage people to come and visit, and I did. Not shaming them, but just saying, “[[Spouse]] would love a visit. Any chance of you stopping by this week?” So, I tried to do as much of that as possible. (DPC c16, bereaved care partner, male, 61, non-AD)
	B. Care partner facilitating group socializing: it’s something that aggravates him when we’re in a group of people, and he’s—everybody’s having a conversation, and he’s not in it, not that he could be. But he feels like people think he’s a freak or something like that. So what works best, and we do have some friends who are really patient with him, is when he’s one-on-one with somebody. No one else … so I will often… I don’t intentionally arrange it but if, you know, a couple comes and I’ll just excuse myself and go do something in the kitchen or something, and they’re very comfortable talking with him. We have some really kind friends. (DPC c19, care partner, female, 72, AD MMSE 17)
	C. Care partner maintaining social network: “You can’t just expect that people understand what you’re going through, because they don’t. … And so I’ve figured out that if I have to—if I need something, and a friend could really help out … I mean, really do have to get some help. I’m not going to be able to do this alone. So it’s about asking people. (DPC c27, bereaved care partner, female, 68, AD)
Care partners receiving a break	D. Paid or informal support: “The thing that helped the most was really being able to afford having private caregivers for my husband, so he didn’t have to… I wasn’t worried about him constantly at [[facility]]. (DPC c16, bereaved care partner, male, 61, non-AD) E. Respite programs: [A program] had a weekend respite thing. It was the first time I had three days… well, it was basically two days of freedom. (DPC c04, bereaved care partner, male, 67, non-AD) F. Respite program: Interviewer: When you were in the respite home, when you were at the VA for a week, how was that? Person Living with Dementia: Oh that was horrible. They were sort of demanding, the nurses were. it was pretty intense … Caregiver: Yeah … there were other people in the room but they were curtained off so he really had privacy … . Interviewer: And what did you do with your time? Caregiver: Oh my granddaughter was home for that week so we went off first thing and got massages. < laughs > And then the next day we went and got manis and pedis, right, and after that went to the Hotel and had a drink and looked at the view and then there was something else that we did together but it was a good week. (MDC c20, dyad; P020, person living with dementia, male, 87; c020, care partner, female, 83, AD, FAST Stage 6 (moderate))
Shared social activities	G. Support groups: “it was caregivers and the patients and then they’d meet together and then we’d separate into two groups … from that, a good core group of six couples … have continued to see each other, have dinners at our houses … it has been kind of ongoing and helpful and good friends. (DPC c27, bereaved caregiver, female, 68, AD) H. Support groups: “Well, other than the medical treatments I think going to the support groups for both [[Wife]] and me has been important.” (DPC p04, person living with dementia, male, 73, AD, MMSE 27)
	I. Maintaining a range of activities: “we try to keep busy … we went on a hike last Saturday, in [[State Park]]. During the fall and winter and spring we take … at the … state university here … classes … last class we took was dealing with Motown … The Supremes, Temptations, Four Tops, so on, and then we took a class on the Supreme Court … . I think she enjoys it. I enjoy it. … it’s just—it’s fun. (DPC c30, care partner, male, 67, non-AD)
	J. Grandchildren: Instead of socializing with our friends we go down and see the grandkids …It’s not entirely just because of [[Husband]]. It’s where we’re at at this stage of our lives. (DPC c29, care partner, female, 64, non-AD, MMSE 24) K. Grandchildren: Person Living with Dementia: Since I don’t have a wife, so the most important people in my life are my son and my daughter both, the two people most important, see, in my life. Son: How about your grandkids? Person Living with Dementia: Oh, yeah. Okay. < laughter > Yeah, my daughter’s and son’s kids. … Son: Grandkids. Person Living with Dementia: Grandkids. (MADC d08, dyad; p08, male, 82; c08, male, 48, AD, FAST Stage 5 (mild))
	L. Music and art: Daughter: We fold laundry. We listen to music … She loves Frank Sinatra. She had a CD player and it was old, so I bought a new one the day before yesterday … when I was in Target I just happened to find a Frank Sinatra CD. She loves Frank Sinatra … And she’ll sing-along, if she knows the song she’ll sing-along. (MADC c10, care partner, female, AD, FAST Stage 6 (moderate))

Notes: AD = Alzheimer’s dementia; DPC = Dementia Palliative Care Study; MDC = Music and Dementia Caregiving Study; MMSE = Mini-Mental Status Exam; MoCA = Montreal Cognitive Assessment; non-AD = non-Alzheimer’s dementia (includes dementia conditions listed in Table 1).

When we’re in a group of people, and … everybody’s having a conversation, and he’s not in it … he feels like people think he’s a freak or something like that. So what works best… is when he’s one-on-one with somebody. No one else… I don’t intentionally arrange it but if… a couple comes, I’ll just excuse myself and go do something in the kitchen or something, and they’re very comfortable talking with him. We have some really kind friends. (DPC c19, care partner, female, 72, AD, MMSE 17 points)

In some cases, external support services provided much-needed breaks in caregiving. For example, paid supportive services (Quote 4D) and use of respite programs (Quote 4E) provided meaningful time for care partners to reconnect with social networks (Quote 4F).

In addition, care partners described developing or adapting shared social activities to preserve their relationship quality at various stages of the disease trajectory. For example, early on, support groups for both care partners and persons with dementia provided an expanded social network who understood the challenges of dementia (Quotes 4G and 4H); participants described meeting with other members outside of the formal support groups for dinner and other activities without fear of embarrassment or exclusion. Care partners developed a broad range of potential shared activities for different stages of dementia (Quote 4I), including hikes, concerts, spending time with grandchildren (Quotes 4J and 4K), or music engagement (Quote 4L).

Discussion and Implications

This multisite qualitative study provides an in-depth examination of the social lives of persons living with dementia and care partners at different stages of the disease trajectory. The sample showed substantial heterogeneity in social experiences by dementia type, patterns of progressive cognitive loss, and individual experiences. Yet, there were several unifying features across the sample. Findings indicated that dementia, and disease-related stigma or social exclusion, was a major disruptive force in the shared social life of persons with dementia and care partners. Social changes generated profound loneliness and detrimentally affected quality of life for both members of the dyad. Participants further described that maintaining social well-being was a moving target; coping strategies and shared social activities that seemed adequate would quickly lose relevance with accumulating or intersecting cognitive deficits, caregiving responsibilities, or changes in social network structures. Consequently, care partners described mixed experiences of social withdrawal, creativity, and adjusting their expectations as they attempted to meet the unique needs of themselves and persons with dementia.

A large body of literature describes the quality of life and both negative and positive aspects of dementia-related caregiving (Cross et al., 2018; Lloyd et al., 2016; Quinn & Toms, 2019). We build on prior literature, adapting the IBM conceptual framework to show findings regarding the unique needs of persons with dementia and care partners across heterogeneous disease stages and syndromes (Figure 1). In this model, the individual social lives and combined “Social Capital of the Dyad” is negatively affected by dementia. For persons with dementia, each cognitive deficit generated a barrier to a previously enjoyable activity with external social networks; memory loss impaired spending time with family, executive dysfunction prevented independence in safe transit to community groups, and apathy reduced one’s interest in all activities. With progression of disease stage and cognitive decline, dysfunction in multiple cognitive domains was challenging for persons with dementia and/or care partners to overcome and resulted in social exclusion from the surrounding network of friends and family. In addition to disease-related barriers, emotional discomfort (e.g., shame, disease-related stigma, feeling burdensome) within dyads and within surrounding social networks led to loneliness and social isolation in persons with dementia and care partners alike. Memory loss in the person living with dementia, for example, generated embarrassment, apathy among persons with dementia created guilt among care partners, and care partners perceived that the person living with dementia’s behavioral changes upset others. Coping strategies required continuous adaptation and creativity, typically through care partners facilitating social interaction with external networks, shared social activities, or paid/external support services.

In addition, our model highlights the core of experience of loneliness for care partners due to changing relationship roles along the dementia trajectory. Children, spouses, and friends described a loss of intimacy with the person with dementia and missed the prior roles that both they and persons with dementia used to have. Consistent with prior literature, events such as a person with dementia moving to a care facility were, for care partners, accompanied with mixed feelings of relief from caregiving burdens and profound sadness due to the loss of companionship (Hennings & Froggatt, 2019; Mullin et al., 2013). These changes were a constant reminder of the progression of the disease. Participants described anticipatory grief around the losses of relationships and circumstances that used to be, along with the losses to come, which could itself precipitate loneliness and social isolation.

The need to enhance social connection to improve quality of life and health has received recent attention since the COVID-19 pandemic, including in a 2020 National Academy of Sciences Report and the 2023 Surgeon General Report (National Academies of Sciences, Engineering, and Medicine et al., 2020; Office of the Surgeon General, 2023). Yet, evidence surrounding interventions for persons with dementia and care partners in this space is lacking. Our results point to potential mechanisms to test in interventions to promote social connection among persons with dementia and care partners. First, results suggest interventions and clinicians should screen for loneliness and isolation among persons with dementia and care partners. Standard methods that clinicians use to assess loneliness and isolation exclude persons with dementia who have difficulty answering subjective scales. Clinicians often do not assess care partner’s experiences of loneliness and isolation (or well-being). The lack of screening may in part be because under most common payment models, clinicians of persons with dementia cannot be reimbursed for caring for the care partner, and care partners often report avoiding seeking healthcare for themselves due to the time, effort, and money being spent on caregiving. Future work is, therefore, needed to improve and test clinical assessment scales (e.g., as has been done for subjective pain scales) for persons with dementia and processes to systematically screen for loneliness and isolation among dyads affected by dementia across the disease trajectory. Notably, interviews in our study reveal that broad, open-ended questions about challenges and needs presented by an empathetic interviewer elicited discussions involving loneliness, social isolation, and social needs (Supplementary Material Section 1), which may be applied to clinical settings.

Second, interventions should be studied that consider how to reduce social isolation and loneliness of the dyad through shared social activities. Given the central role of care partners in facilitating social activities as indicated in our study and updated IBM model, interventions should account for the capacity for care partners to be involved. For example, support groups were described as helpful because each member of the dyad received support, and care partners were under less social pressure as they could develop a network of people already familiar with the impacts of dementia. persons with dementia can likely participate in many existing interventions for social isolation and loneliness (e.g., intergenerational engagement, group classes, outdoor activities/gardening, social robots; Poscia et al., 2018) depending on the pattern of cognitive decline. Interventions can be further adapted to the needs of persons with dementia–care partner dyads through creating comfortable spaces and stigma reduction, an area of active investigation (Kim et al., 2019). Prior research indicates that persons with dementia can engage in adapted meaningful activities throughout the disease trajectory (Allison et al., 2022; Oh et al., 2021). Clinicians might proactively brainstorm with care partners and persons with dementia (when possible) to adapt previously enjoyable activities and account for individual cognitive and personal needs; such approaches are being explored in coaching interventions (Van Orden et al., 2023). Other examples of shared activities include music or art programs, even when persons with dementia have memory, language, or visuospatial impairment (Johnson et al., 2020). Virtual support groups may enable participation for care partners or in rural locations (Henderson et al., 2022; Velloze et al., 2022). Technologies (e.g., Virtual Reality devices, robotic pets, and others) may have a role in facilitating social interaction or companionship (Appel et al., 2021).

Third, although many interventions to reduce loneliness among care partners focus on behavioral activation (i.e., Self-Determination Theory; Ryan & Deci, 2014; Van Orden & Heffner, 2022; Van Orden et al., 2023) researchers and clinicians should acknowledge the profound and persistent feeling of loss of companionship, documented throughout Theme 2, that contributes to long-term loneliness and grief. Support for care partners might, therefore, draw on interventions designed to address pre- and post-death grief earlier in the disease course (typically designed for complicated grief, or what the DMS-V now calls prolonged grief disorder; Prigerson et al., 2022; Rubin et al., 2022). Other interventions might involve psychologists and dementia-trained social workers who focus less on “fixing” loneliness or grief and more on validation, anticipatory guidance, and longitudinal support (Bannon et al., 2022; Dehpour & Koffman, 2023; Kasl-Godley & Gatz, 2000). Interventions may need to be proactive, flexible, and easily accessible since care partners may be overwhelmed or struggle with time or motivation to participate in these programs.

Fourth, results add to prior evidence that external programs and policies that reduce caregiver burden may increase opportunities for persons with dementia and care partners to remain engaged in meaningful social activities (Allison et al., 2022). Currently, services tend to focus on more advanced disease and relieving physical caregiver responsibilities such as hospice and respite programs. More flexible policies of support (including payment mechanisms) that enhance social connections throughout the dementia trajectory will be needed to preserve and prevent social loss instead of current approaches which narrowly focus on physical caregiver needs or too-late attempts to recover lost social networks later in the disease process or bereavement stages (Poscia et al., 2018). Recent initiatives by the Centers for Medicare and Medicaid Services to develop methods for health systems to receive reimbursement for screening for health-related psychosocial needs of dyads, including needs of unpaid caregivers, have substantial potential to improve dementia care quality (Guiding an Improved Dementia Experience [GUIDE] Model, 2023).

Study findings should be interpreted in the context of limitations and study considerations. As this is a secondary data analysis of qualitative interviews, we were not able to return to participants to conduct additional interviews or member reflection to clarify thematic findings around social changes. Although changing social lives did emerge among persons with dementia, they were disproportionately discussed by care partners, particularly at more advanced stages of disease. In addition, interviews were conducted prior to the pandemic, which profoundly altered social lives. The same study conducted today may uncover even greater loneliness, isolation, and unmet social needs. In addition, the DPC study was relatively homogenous by race and ethnicity, the MDC study was relatively diverse (47% of persons living with dementia and 59% of care partners were people of color), and both studies had on average a high education level. Future research incorporating diverse participants, including Hispanic/Latinx, non-English-speaking participants, and persons of varying socioeconomic status is needed given the differing social and cultural experiences of dementia.

In conclusion, this qualitative study of persons with dementia and care partners identified social isolation and loneliness as a shared experience and source of distress throughout the disease trajectory, including bereavement for care partners. Interventions that can reduce the intensity and breadth of social losses experienced by persons with dementia and care partners are likely to improve the quality of life. These efforts should be individualized to address the needs of dyads as changes and crisis points occur along the disease trajectory.

Funding

This work was supported by the National Institute on Aging [K01AG059831 to K.L. Harrison, K23AG065438, P30AG044281 to A.A. Kotwal, R01AG057751, K24AG068312 to A.K. Smith, K23AG062613 to T.A. Allison]; the National Institutes of Health [KL2TR001870 to K.L. Harrison].; the National Palliative Care Research Center [Kornfield Scholar Award to A.A. Kotwal]; and the Hellman Foundation Award [Early-Career Faculty Award to A.A. Kotwal].

Conflict of Interest

A.A. Kotwal and C.M. Perissinotto report grants from Humana Inc. outside the submitted work and research consulting fees from Papa Health outside the submitted work. All other authors report no potential conflicts of interest.

Data Availability

Data are not available due to issues with blinding the data to protect privacy of participants. Studies reported in this manuscript were not preregistered.

Acknowledgments

We are grateful to Jennie Gubner, PhD, and Annette Rodriguez, MA, for their role in MDC study interview guide development, data collection and code book development.