Abstract
Introduction
Loneliness is generally understood as the discrepancy between an individual’s preferred and actual level of social contact. According to a new report from the National Academy of Sciences, 43% of adults aged 60 and older report loneliness, while 24% are considered socially isolated (National Academies of Sciences, Engineering, and Medicine, 2020). Loneliness and social isolation are pervasive for individuals of all life stages, yet older adults may be especially vulnerable due to natural age-related transitions. Social networks of reduced size and diversity, along with less communication with friends and family, increases the risk of loneliness (Anderson & Thayer, 2018). In addition, personal factors such as physical or cognitive decline, sensory deprivation, retirement, bereavement, caregiver burden, and strained family relations may lead to loneliness and social isolation (Donovan & Blazer, 2020; Seeman, 2000).
Loneliness may adversely influence the individual’s beliefs about social interactions. For example, they may experience increased hypervigilance for social threats and form negative biases that may further amplify feelings of loneliness (J. T. Cacioppo & Hawkley, 2009). Among individuals age 60+, loneliness is a predictor of decline in adaptive skills (Perissinotto et al., 2012) and adversely influences mental and physical health outcomes, including depression, quality of life, health utilization, and mortality (J. T. Cacioppo et al., 2006; Luo et al., 2012; Musich et al., 2015). Socially isolated individuals are at an increased risk of adverse outcomes as well, including cognitive decline (Bassuk et al., 1999), cardiovascular disease (Barth et al., 2010), and mortality risk (Eng et al., 2002; Heffner et al., 2011; Kaplan et al., 1988).
Interventions often seek to change loneliness to improve wellness and psychological well-being (Bartlett & Arpin, 2019; Krause-Parello et al., 2019; Schoenmakers et al., 2012). Cacioppo’s classification of loneliness-reducing interventions offered four primary categories: (1) increase social contact; (2) improve social contact; (3) enhance social skills; and (4) address the negative social beliefs or biases (S. Cacioppo et al., 2015). Among these categories, interventions intended to address the social cognitions were associated with the largest effect size. Thus, it is reasonable to infer that interventions using multi-pronged approaches, particularly those offering cognitive behavioral therapy (CBT), may prove efficacious in reducing loneliness.
Increased loneliness resulting from social and physical distancing restrictions amid the COVID-19 pandemic reinforces the benefit of providing web-accessible content. The advent of the pandemic underscored the potential benefit of this type of intervention, particularly for older adult populations who are considered higher risk for complications and deaths related to COVID-19 (Centers for Disease Control and Prevention, 2021; Luchetti et al., 2020). Web-based interventions are cost-effective and well-suited for older adults, whose social networks are often geographically inaccessible, especially during the pandemic while many older adults remained at home amid social distancing efforts (Aalbers et al., 2011; Ajrouch et al., 2001). Use of technology to facilitate virtual communication is increasingly recommended for lonely and socially isolated individuals (Berg-Weger & Morley, 2020). In fact, a review of technology adoption among older adults demonstrated that technology use within this population has been increasing although with prevalence and type of use varying by personal characteristics (MacLeod et al., 2019). There are many benefits for older adults adopting new technologies, including that the information can be presented in simple formats (e.g., videos, graphics, audio), it can reach people in their homes and in rural areas, and the anonymity of the online format reduces stigma related to seeking treatment.
A recent review of loneliness interventions for older adults explored approaches that might hold promise for this population, with the majority of interventions classified into one of four categories, including online and digital solutions (MacLeod et al., 2018). Researchers describe social media websites and applications designed to target older populations, along with interactive home devices and “pets” that allow communication with the user. For instance, smart speakers or home assistants often have the capability to respond to voice commands, answer questions, and speak to the user with a digital voice, providing the sense of a social “connection.” Further, telehealth visits for mental and physical health have also emerged to provide needed services as well as a social connection for vulnerable older adults, even more so during the pandemic. However, some options may require existing technological skills or detailed training, especially for those who may not have had sufficient exposure to emerging technologies.
More recently, reviews have explored the utilization of remotely delivered interventions to address loneliness among older adults considering the magnified importance of these options during the pandemic (Gorenko et al., 2021; Ibarra et al., 2020). Gorenko et al. (2021) reviewed the methodology and outcomes of remote approaches to reduce loneliness through video or telephone conferencing; social media applications; peer befriending; and self-guided or therapist-led telehealth interventions integrating CBT, problem-solving therapy, or life review. The authors outlined the characteristics of interventions most likely to produce positive outcomes: active participant engagement; flexible implementation; and integration of older adults’ perspectives in development. Notably, remote interventions have recorded positive outcomes including reduced loneliness, reduced anxiety and depression, increased social connections, and increased self-confidence (Ballantyne et al., 2010; Cattan et al., 2011; Newall & Menec, 2015; Tsai & Tsai, 2011; Tsai et al., 2010).
Nevertheless, a number of inherent barriers compromise the potential success of remote interventions targeting older populations, such as attitudes toward technology; limited experience and technological skills; access to tools; and need for assistance. Thus, development of remote interventions must consider solutions to these barriers prior to implementation. Consequently, there are certain gaps in the availability of remotely delivered interventions for loneliness among older adults, with most existing approaches facing certain challenges. The pandemic crisis brought additional urgency to the need for remote interventions, with social distancing and daily routine upheaval compounding loneliness for many older adults. In the current study, we explored the experiences of older adults who participated in a web-based intervention intended to alleviate loneliness. Findings from this study will contribute to existing literature by examining what, if any, actionable steps participants take to alleviate loneliness as a result of participation.
Methods
This study was the second phase of a larger multi-phase research study intended to better understand the health-related issues of older adults covered by AARP® Medicare Supplement plan. The AARP’s Aging Strong initiative intends to improve insureds’ personal and social investments in well-being through a series of interventions aiming to increase resilience through enhanced purpose in life, social connectedness, and optimism (Yeh et al., 2019). This study was approved by the New England Institutional Ethics Review Board, an independent institution that reviews protocols for nonacademic institutions.
Participants
AARP Medicare Supplement insured participants were recruited for individual interviews during the first phase of the study. Inclusion criteria consisted of those who previously identified as being lonely and/or having low resilience using a screener that included the UCLA-3 and Brief Resilience Scale (Connor & Davidson, 2003) administered via interactive voice response (IVR) survey in conjunction with AARP’s Aging Strong initiative. Exclusion criteria included not enrolled in an AARP Medicare Supplement plan, currently engaged in care coordination, younger than 65, on the “do not call” list, no valid phone number, and not web enabled.
Eligible participants from the first phase of the study were stratified according to age range and gender, with the aim of recruiting an equal proportion from each segment. Following recommendations for a sample size of 12 to 20 participants in an interview study (Lincoln & Guba, 1985), investigators planned an initial goal of 20 interviews, with subsequent assessment of whether data saturation had been achieved in order to recruit additional participants if necessary (Francis et al., 2010).
Overview of the Intervention
The intervention was developed by doctoral-level psychologists with experience in evidence-based programming and loneliness. Cognitive-behavioral, acceptance, mindfulness, and positive psychology concepts, traditionally delivered via face-to-face, were adjusted into easy-to-understand digital materials that were user-friendly for older adults.
The content was presented in eight modules, each with a lesson, a quiz, and an activity designed to apply the content. In addition, a designated prompt at the end of each lesson directed participants to schedule or initiate a chat with a coach to further discuss application of the content in their daily lives and develop SMART goals (specific, measurable, actionable, relevant, timely).
Session modules included the following: (1) Values and social connectedness, (2) Coping, (3) Awareness of thoughts, (4) Impact of thoughts, (5) Changing Impact of Thoughts (labeling and challenging), (6) Changing Impact of Thoughts (acceptance and self-compassion), (7) Relationship skills: social stuck points, and (8) Preparing for setbacks and having a plan.
After completing each session, participants were “locked out” and prevented from accessing future module content for 7 days to encourage further contemplation. After participants voiced frustration about the lockout duration, time was reduced from 7 to 4 days. Overall, 28 of 72 participants finished the program before the change in lockout duration. On average, those who completed the program (n = 73) took 78 days (11 weeks).
Participants used an embedded chat messenger to initiate a one-on-one session with a trained coach after each module. Program content advised participants they should only share content they were comfortable disclosing to the coach. Coaches, only identified by an avatar and fictitious name, were individuals with backgrounds in the behavioral health field. To become initial coaches, they were required to go through several rounds of training centered on learning about the platform, effectively interacting with users, CBT, and crisis protocols. For the trial, we conducted classroom-like sessions in-person or via video calls to conduct training specific to the loneliness program and to go over individual lessons and activities, chat goals, and program structure. All coaches were then required to go through the program to gain a user’s experience, and were provided a chat summary syllabus for each conversation to guide the call.
When interacting with live users, coaches were provided information on the user (e.g., age and gender), transcripts of their previous coaching interactions, as well as a checklist of tips outlining the high-level topics/goals for that interaction. Post-interaction, conversations were submitted for administrative review to ensure the coaches met the objectives of the chat, and the coaches were provided feedback. Similarly, coaches met periodically with administrators and other coaches to review user interactions and share lessons learned to strengthen their skills and improve overall interactions throughout program delivery.
Data Collection
A marketing research company received the full list of eligible participants (n = 73) who enrolled in the program. Recruiters were instructed to recruit evenly among age range, gender, and completion status, given the available sample. Recruiters contacted participants by telephone, verified identity, explained the study, and scheduled interviews with the first 20 recruits. Verbal consent was obtained prior to the start of the interview. Interviews lasted approximately 1 hr; no personal identifiers were collected, and all interviews were audio recorded and transcribed verbatim.
The interview guide consisted of 13 questions, which elicited feedback about initial impressions of the program, motivation for joining, personal comfort with online technology, reactions to the organization of the website, potential goals, anticipated outcomes, SMART goals, helpful content, perceptions of the program’s “fit” given their needs, learnings derived about themselves, whether they acquired any new problem-solving skills, and the outcome of program-related goals.
Analysis
Investigators analyzed participants’ transcribed interviews using qualitative description. Qualitative description was an ideal methodology for this data as it draws from a naturalistic perspective, offers flexibility in commitment to a theory, involves review of interview data, and allows for maximum variation sampling (Kim et al., 2017).
Two investigators (JH and RU) conducted a qualitative content analysis using an iterative, constant comparison process. During the coding process, both coders independently read transcripts, identified an initial code list, and developed operational definitions. Then, coders conducted line-by-line coding that included comparison and refinement of identified coding between both investigators. Coders subsequently discussed, reviewed, and reread interview data to develop final coding and reach consensus about meaning (Ryan et al., 2000). One investigator coded all transcripts while the other coded 50% of the overlap. Both investigators reviewed coding on overlapping transcripts to reevaluate passages coded across researchers, and the codes applied based on the assigned definition in the codebook (Creswell & Poth, 2017). Any conflicts in assigned codes were settled through discussion until consensus was reached.
Next, both investigators examined the properties and categories of all codes to identify opportunities for categorization according to shared properties. Investigators subsequently used this categorization of codes to develop overarching themes that described patterns of usage and provided a narrative of participants’ overall use. Final coding was imported into Nvivo (2018) a qualitative software program.
Results
Twenty-four individuals participated in the study, with a slightly higher distribution of males (n = 13) than females (n = 11). Breakdown in age range was as follows: 65 to 69 (n = 9), 70 to 79 (n = 13), and 80 to 89 (n = 2). The average participant age was 73.
Dosage data showed a mean of 62.2 days participating in the program (from the first day to the last). Participation in this study was higher than that of the overall population of 151 participants at a mean of 58 days. Dropoff for participants was measured using the following designations: completed all eight program modules and the T2 survey (n = 14), completed five to seven modules (n = 6), and completed fewer than four modules (n = 4).
Investigators developed three themes that reflected participants’ experiences, feedback, and reported outcomes: Program Goals, Experiences with Course Engagement, and Program Outcomes.
Program Goals
When considering participation in the program, participants described varying motivations. Several members wanted to improve their mental health, including the few members who described having a history of anxiety or depression (n = 3) and previously participating in therapy (n = 3). Participants also described feeling lonely, socially isolated, and struggling with recent transitions, such as the death of a spouse, retirement, or relocating (n = 6).
I’m a widowed woman and I’m home by myself most of the time. My son lives with me, but I’m 78, he’s 60. We kind of live in our own worlds and I didn’t realize that after his daddy died, I’d slip that far into depression.
My wife had died and it seemed the purpose of the therapy of program was to make a social adjustment and get me on track again.
Many participants hoped to improve their social skills, including overcoming social anxiety (n = 11). Some reported feeling anxious and doubtful of their ability to initiate social contact or confidently socialize with new acquaintances (n = 5). Participants also described negative self-talk about their socialization skills, including the belief that others would assess them negatively in new social settings (n = 4):
I never talked a whole lot because I always thought that everybody was critical towards me because I have my own opinions and it’s pretty hard to change.
These members hoped they could acquire new skills and strategies for initiating and holding sustained conversations, accepting invitations to socialize, and feeling comfortable in group settings. A few members hoped the program would equip them with skills to improve the quality of existing relationships (n = 3).
Perhaps just to be more patient – I’m one of those quick-tempered people. Now that I’ve gotten older, I’ve mellowed a lot. But there’s room for improvement.
Accordingly, several participants reported setting SMART goals centered around setting up social encounters, initiating outings, joining social clubs, and attending social opportunities affiliated with their synagogue or church (n = 8). Participants also reported SMART goals related to consistently communicating with friends and family members at regular intervals, improving existing relationships and reconnecting with long-lost acquaintances (n = 5).
Experiences With Course Engagement and Rapport Building
Overall, most participants found the program content and related activities helpful (n = 15). Several participants explained that the program’s process flow of didactics followed by reinforcement exercises and chats with coaches to explore “real-life” applicability helped spur action from contemplation to behavior change (n = 7). Participants perceived the process of goal-setting beneficial, coupled with the accountability of a coach. In several instances, participants recounted setting specific goals related to increasing social interaction and reported back on the success of the strategy with their coaches (n = 5).
When asked to describe the content topics they found most helpful, many cited topics that addressed negative self-talk (n = 8). Similarly, participants found program content related to self-compassion and self-acceptance to be helpful in overcoming their anxiety over social situations (n = 6).
There was one particular lesson that taught us how to recognize the little voice in our head, and how to turn it off and go beyond it and express yourself beyond it. I thought that helped because usually I am the one who says, ‘Oh, they’re not interested in me. Why should I bother?’
What this program did was take a little bit of fear out of some of these challenges by thinking in the back of your mind of acceptance and self-compassion.
Many participants described receiving helpful suggestions and advice from their coaches (n = 14). Coaching feedback deemed most helpful centered around providing practical tools and strategies for making new friends and acquaintances, including coping with social anxiety (n = 8), navigating personal relationships (n = 3), and managing conflict with family members (n = 2). Participants paired with coaches who utilized rapport-building techniques were more likely to report being satisfied with their interactions and more receptive to the advice and/or suggestions offered by coaches (n = 10). Overall, participants found coaching most beneficial when the coach provided personalized feedback as opposed to standard responses, elicited feedback about the user’s life, and took time to fully address any concerns before ending the interaction (n = 10).
She would ask thought-provoking questions about the previous week, what I had done, did I have any issues that had cropped up. And she was able to ask some questions that I might not have thought of, so that made me dig a little deeper. I think that is definitely a beneficial part of the program.
Conversely, participants who perceived their coaches as “distant,” or focused primarily on administering program content without making an effort to learn more about their daily lives, were more likely to be dissatisfied with their coaching experience (n = 8). Participants were disinclined to continue interaction with coaches who failed to attempt rapport building (n = 7), expressing displeasure when coaches rarely deviated what from what was perceived as an outline. These participants were often more likely to report a preference for having one principal coach who followed their progress and with whom they could form a close rapport. A few members who were switched from a long-term coach reported difficulty acclimating to new coaches with approaches that did not seem warm (n = 5).
I thought I was talking to a machine rather than a person. I would present a situation to the coach. The answer was, ‘Well, you certainly seem to have that under control.’ But when I talked to another coach and came back to a similar situation, they’d say, ‘Well, its seems you’ve got that concept under control.’ And I was talking to two different people. I didn’t feel too good after that. [Dropout, 5–7 sessions]
My next coach was vague. And when I would tell her things I told the first coach, she was very vague in her assessment of what I saying. She just didn’t seem as interested and she didn’t know how to connect what I was telling her.
Perceived Program Outcomes
Participants reported numerous positive outcomes related to increasing their social activity and alleviating loneliness (n = 17). Many participants reported their participation and goal-setting led them to make a conscious effort to identify and initiate opportunities for social interaction (n = 11). This often consisted of identifying daily opportunities to connect or reconnect with friends and acquaintances, taking initiative to create new social groups, and committing to future goal setting around social interaction. Participants reported achieving goals such as making a deliberate effort to greet others during errands or when entering unfamiliar social settings and identifying opportunities to connect with friends and acquaintances. One user shared: I said I was going to rejoin the Soaring Club. I was going to start flying gliders again. So I’m flying gliders. Other examples include attending book clubs, establishing new routines with others at church and synagogue, and reconnecting with old friends.
For many, the importance of proactively setting future goals around social interactions was a key learning (n = 7). Accordingly, these participants noted their plan to continue setting goals after the conclusion of their involvement in the program.
I say, ‘I’m going to do that. I’m going to meet with three people and go out to lunch. . .’ I find myself doing that now. (5-7 sessions completed)
I constructed a group of people that I thought had common interests and developed a golf group, which is still ongoing today, so I think it worked out pretty well.
Several participants reported progress addressing self-perceived awkwardness or lack of confidence related to interpersonal communication reported improvement (n = 8). Program content related to “sticky thoughts,” “all or nothing thinking,” and self-compassion helped these participants identify and neutralize critical self-talk impeding their efforts to connect with others. For example, one male participant commented: I don’t beat myself up like I used to over small stuff. Another participant explained:
I think I have tendency to judge myself quite harshly. And so I feel the program helped me see that I’m not the only person on the planet that’s struggling with these interpersonal issues. And if I’m struggling, there’s ways for me to circumvent it, to help alleviate it and get rid of it all together.
Participants also reported gains in managing difficult interpersonal relationships (n = 13). Program didactics and individual coaching provided tools for participants to facilitate better coping with conflict. These participants described changes to their own behavior that improved their relationships, including managing anger, monitoring defensive behavior, and setting healthy boundaries (n = 13).
I’m a little more conscious of people’s feelings and how to approach them and being a little less aggressive. Now I give people the chance to talk or answer fully or being patient to wait for an answer, or to solicit an answer instead of just saying this is the way it is.
It made me look at stuff more open-minded instead of a quick response to that particular need. I now think about it a little more. How should I answer it and how will I answer it? Because ‘should’ and ‘will’ are two different things and sometimes my mind should be saying, ‘Let me think first and then I’ll say this is what I want to do r suggest we do. So it’s a two-way street and it makes you think, which I appreciate.
Discussion
Results from the current study show older adult participants benefited from the multi-pronged approach of this web-based intervention, with potential amelioration of negative biases toward social interaction and increased social contact. These findings align with the RAM framework and show the benefit of intervening on prolonged loneliness by addressing hypervigilance and social monitoring. Consistent with previous findings, participants’ negative self-appraisals of their social skills initially impeded their willingness to attempt new social connections. Program didactics with underlying mindfulness tenets, coupled with reinforcement from coaches, addressed internal barriers to participants’ willingness to engage with others, and provided practical tools for increasing social contact. Further, participants reported improvement in the quality of their personal relationships and greater confidence with initiating connections. This feedback suggests the intervention potentially meets S. Cacioppo et al.’s (2015) criteria for assessing the efficacy of loneliness-reducing interventions. Further, results shown here confirm previous findings in loneliness interventions integrating telephone discussions, video calls, internet-based education programs, and befriending programs, in which participants have reported reduced loneliness and depression along with increased social connections; emotional and social support; and confidence (Ballantyne et al., 2010; Cattan et al., 2011; Newall & Menec, 2015; Tsai & Tsai, 2011; Tsai et al., 2010).
Program didactics related to abandoning “all or nothing thinking” and encouraging self-acceptance or self-compassion were considered particularly helpful. Content equipped users with resources to confront and resolve anxiety about new acquaintances’ potential assessment, and about their own ability to interact with others. Participants reported these tools were effective in identifying and processing self-defeating and negative self-talk, which facilitated their efforts to attempt new social contexts. The acceptance and mindfulness philosophies on which this intervention was based encourage participants to accept, rather than resist, their anxiety over social interactions. This approach may be ideal when addressing loneliness among older adults, particularly those with depression and anxiety. Individuals with depression or anxiety report greater social isolation when compared to those without depression or anxiety, even when there is no difference in social contact. This may be attributed to a tendency for depressed individuals to hold more negative social expectations, and to have more of a negative bias toward social interactions as compared to those without depression. In previous studies, similar approaches have resulted in reduced depression, reduced loneliness, and increased confidence following befriending activities that boosted self-acceptance (Cattan et al., 2011).
While this intervention appears to provide promising outcomes, participant feedback indicated a preference for personalized coaching feedback. For example, participants more positively regarded coaches who approached interactions with rapport-building techniques such as small talk. Coaching recommendations and goal-setting support most closely resonated with participants when the coach offered specific, tailored recommendations. This finding illustrates users’ desire to connect with assigned coaches, congruent with earlier findings that older adults considered personal, one-on-one support crucial to their experience in internet and social media training programs (Ballantyne et al., 2010).
Peer coaches may also satisfy participants’ desire for continued rapport-building throughout the coaching process. Further, participants with a history of depression and anxiety may derive more benefit from coaching offered by a credentialed therapist. Future research should examine user acceptance of this model with coaching offered by a professional health or peer coach. Finally, this study was conducted prior to the advent of the COVID-19 pandemic, and as such, the curriculum design encouraged increased social contact using both virtual and face-to-face modalities. Future iterations of this study should adopt a teleconferencing feature to not only connect coaches with users, but also provide users with a platform to virtually connect with others.
Limitations of this study include higher program dosage as compared to participants overall, which may represent a sample bias. Additionally, inclusion criteria for participation in this study included low resilience, which may not have rendered each participant an ideal fit, particularly if they were not lonely and/or socially isolated. The cross-sectional approach of this study does not provide insight into long-term efficacy, including participants’ success in building and maintaining social networks over the long-term and during crises such as the COVID-19 pandemic. Finally, there was no control group in this study, which precludes us from being able to identify which aspects of the approach most benefited the participants.
Conclusion
The current study demonstrated potential efficacy of a web-based loneliness intervention and yielded important findings about older adult participants’ preferred intervention characteristics. Notably, participant feedback showed users desired rapport-building communication when interacting with coaches, as well as personalized feedback that accommodated their personal contexts. Future studies should examine the efficacy of using credentialed therapists, or peer coaches with similar life stage experiences. Given the recent worsening epidemic of loneliness among older adults during the COVID-19 pandemic, future iterations of this model may emphasize facilitating social connection using virtual platforms such as videoconferencing. Finally, longitudinal studies are needed to determine the long-term efficacy of participants’ newly forged and strengthened social connections over time.
Footnotes
Acknowledgements
We thank Stephanie Macleod for her careful review of the manuscript.
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.
Ethical Statement
All subjects gave their informed consent for inclusion before they participated in the study. The study was conducted in accordance with the Declaration of Helsinki, and the protocol was approved by the New England Institutional Review Board (NEIRB #120180244), an independent institution that reviews protocols for nonacademic institutions.
HIPAA Identifiers
No participant identifiers are included in this manuscript.
