Social interactions significantly impact the quality of life of adults in general and older adults in particular. Health risks have been associated with the characteristics of each individual’s social network, such as small size [
Several studies report that both loneliness—a subjective measure referring to the “unpleasant” lack of (quality of) social relationships [
Indeed, several risk factors such as sensory incapacity and reduced mobility, as well as reductions in the quality and frequency of contact and requirements for long-term care or additional support, are associated with loneliness in old age [
Technological innovations, along with social and economic changes, have made interconnected devices a commonplace, thus creating opportunities for interaction [
In this systematic review, we focus instead on interventions enabling long-distance interactions through technology-mediated communication, targeting loneliness and social isolation in old age. Our objective is to identify the findings and limits of the knowledge acquired so far and to emphasize areas where further research is needed. More specifically, for the interventions analyzed, we investigate the following research questions:
In the following, we discuss our investigation methods and results.
We conducted a systematic review [
We analyzed the title and abstract of each of the 1178 search results and verified whether the publication targeted older adults. For the purpose of this review, we adapted the definition of older adults by the WHO [ The work included an intervention (i.e., action taken to improve a situation) The interactions with people were long-distance The intervention supported mainly technology-mediated communication The impact on loneliness or social isolation was evaluated
Despite our focus on loneliness and social isolation, we considered social connectedness, the experience of belonging and relatedness among people, as a valid outcome because it is related to the (dis)satisfaction with contact quantity and quality [
After the identification and screening phases, 190 publications were left (see Figure
The full selection process, following the PRISMA statement guidelines.
The review follows a narrative approach to the synthesis of results, given the heterogeneity of the studies included. In order to answer our research questions, On the challenges of long-distance interactions addressed by interventions (RQ1), we analyze intervention strategies and outcomes used to accomplish the study goals On the technology used to support interventions (RQ2), we account for the technology and devices, as well as the use of the technology by older adults On the social interactions enabled (RQ3), we describe the different contexts of interaction and the contacts reached by participants
Some authors were contacted to clarify the devices used in their interventions and the strategies that participants used to meet new people online. For two studies [
We start by summarizing the interventions analyzed. There were 13 interventions that considered loneliness as a primary outcome [
Summary of the interventions analyzed.
Intervention | Technology/device | Studya: length/participants/age | Strategies | Study settings/methods | Outcomes: primary/secondary | Measurementsb/Conclusion |
---|---|---|---|---|---|---|
Ballantyne et al. [ | Social network (About my age)/computer | 3 months/4/69–85 | Internet and computer training | Home/pilot study, prepost interviews | Loneliness | Interview/decreased loneliness |
Blažun et al. [ | E-mail, Internet, and Skype/computer | 3 weeks/45/66 in Finland, 77 in Slovenia | Internet and computer training | Home in Finland, residence in Slovenia/prepost test no control; no standard tool for assessment | Loneliness, ICT knowledge, and experience | Questionnaire/decreased loneliness |
Cattan et al. [ | Phone calls/telephone | >3 months/34/55–95 | Familiar, simple technology; ensure interactions | Home/mixed methods | Health and wellbeing/loneliness | Questionnaire and interview/decreased feelings of loneliness, increased socialization |
Cotten et al. [ | E-mail, Internet, and Facebook/computer | 1–2 weeks/205 (79; 126)/82.8 | Internet and computer training | Facility/cross sectional analysis | Loneliness and social isolation/quantity and quality of communications | Hughes 3-items LS (UCLA-based) and questionnaire/decreased loneliness, not social isolation |
Dodge et al. [ | Video chat/touch-screen computer | 6 weeks/83 (41; 42)/80.5 | Familiar, simple technology; ensure interactions; provide conversation topics | Facility/randomized controlled trial | Cognitive function/loneliness | Hughes 3-items LS (UCLA-based)/no difference |
Fokkema and Knipscheer [ | E-mail and Internet/computer | 3 years/26 (12; 14)/66 in intervention, 68 in control | Internet and computer training | Home/interrupted time series, nonequivalent control group prepost test | Loneliness | DeJong 11-items LS and questionnaire/decreased loneliness |
Garattini et al. [ | Broadcast, messages, and calls/touch-screen computer-phone hybrid | 10 weeks/19/65–84 | Familiar, simple technology; provide conversation topics | Home/mixed methods, exploratory study | Feasibility/social connectedness | DeJong 6-items LS, log, interview, questionnaire/helped social connection and created interactions |
Larsson et al. [ | E-mail, Internet, Skype, and Facebook/computer | 3 months/30/61–89 | Internet and computer training | Home/randomized crossover study | Loneliness/satisfaction with social contacts online and offline | UCLA LS/decreased loneliness (significant in both groups); satisfaction with social contacts inconclusive |
Machesney et al. [ | Virtual companion/tablet | 1 week/13/65–93 | Familiar, simple technology; ensure interactions | Home/one group observational study | Loneliness | UCLA LS/decreased loneliness |
Széman et al. [ | E-mail, Internet, Skype, and Facebook/computer | >6 months/15 (program), 25 (pilot)/>75 | Internet and computer training | Home/case study | Loneliness | Observation/increased size of social network |
White et al. [ | E-mail, Internet/computer | 5 months/93 (48; 45)/71 in interventions, 72 in control | Internet and computer training | Facility/randomized controlled trial | Loneliness | UCLA LS (modified anchors)/decreased loneliness (nonstatistically significant) |
Baez et al. [ | Virtual classroom, messages, and predefined messages/tablet | 10 weeks/40 (20; 20)/71.5 | Internet and computer training; familiar, simple technology; ensure interactions | Home/randomized pilot trial | Training adherence/loneliness and social wellbeing | Hughes 3-items LS (UCLA-based)/no significant difference |
Czaja et al. [ | E-mail, Internet, virtual classroom, and messages/computer | 12 months/300 (150; 150)/76.15 | Internet and computer training; familiar, simple technology; provide conversation topics | Home/randomized controlled trial | Loneliness and social isolation/attitude towards technology and proficiency | Hawthorne friendship scale, Cohen perceived social support scale, Lubben social network size, UCLA LS v3/decreased loneliness and social isolation |
Banbury et al. [ | Skype/tablet | 44 weeks/52/73.0 | Familiar, simple technology; ensure interactions; provide conversation topics | Home/nonrandomized noncontrolled prepost test | Educational goals/social support | Social (egocentric) network analysis interviews, focus groups/increased network size |
Barbosa et al. [ | Messages (video, photos, audio, predefined)/tablet | 3 months/12/82.5 | Familiar, simple technology; ensure interactions | Facility/feasibility study | Feasibility/social connectedness | Hughes 3-items LS (UCLA-based), Abbrev. Duke social support index/increased social interactions, high perceived social connectedness |
Chiu and Wu [ | Line (messaging service), YouTube/tablet | 5 months/54 (19; 18; 17)/73.0 | Internet and computer training; provide conversation topics | Facility/group randomized trial | Cognitive, physical functioning and psychological wellbeing/quality of life | CES-D Chinese version, Taiwanese inventory of social supportive behavior/increased social support and satisfaction with contacts |
Gutierrez et al. [ | Video chat, messages, and photos/tablet | 9 weeks/9/69–81 | Familiar, simple technology; ensure interactions | Home/empirical in-the-wild study | Frequency of social interactions | One-way repeated ANOVA/increased social interactions |
Isaacson et al. [ | Virtual classroom, video chat, photos/TV, remote, and webcam | 4-5 weeks/40/85.86 | Familiar, simple technology; ensure interactions | Home/pilot study | Technology adoption/emotional wellbeing | UCLA LS v3, Lubben social network scale/decreased loneliness, increased social wellbeing and social network size |
Jarvis et al. [ | WhatsApp (messaging service)/mobile phone (smartphone) | 3 months/29 (13; 16)/74.93 | Internet and computer training; ensure interactions | Facility/randomized controlled study | Loneliness and social cognition/use of technology | YSQ short form, DeJong 6-items LS/decreased loneliness |
Jarvis et al. [ | WhatsApp (messaging service)/mobile phone (smart phone) | 3 months/32 (15; 17)/70.42 | Internet and computer training | Facility/experimental randomized comparative study | Loneliness | DeJong 6-items LS, focus groups/decreased loneliness |
Morton et al. [ | E-mail, Internet, Facebook, and Skype | 4 months/76 (44; 32)/80.71 | Internet and computer training | Some at home, some at facility/randomized 2 × 2 × 2 study | Cognitive and mental health/social network activity and satisfaction, loneliness | Social network activity index, UCLA LS v3/no difference in loneliness, increased social network activity |
Myhre et al. [ | 8 weeks/41 (14; 13; 14)/80.0 Facebook, 73.38 online diary, 79.29 Waiting list | Internet and computer training; ensure interactions; | Some at home, some at facility/3-arm study | Neuropsychological tests/social engagement | UCLA LS v3, MOS social support survey, Lubben social network 18-i scale/no significant difference | |
Barbosa et al. [ | Video chat, photos, audio recording, predefined messages | 2 months/5/87.2 | Familiar, simple technology | Facility/embedded case study | Feasibility and adoption/social connectedness | Hughes 3-items LS (UCLA-based), Abbrev. Duke social support index/no significant difference |
Pauly et al. [ | E-mail, Internet, social network, and messages | >6 months/92/67.7 | Internet and computer training | Home/prepost, repeated measures study | Physical activity/loneliness and executive functioning | Self-reported questionnaires, R-UCLA LS/no significant difference |
Tomasino et al. [ | Virtual classroom and messages | 8 weeks/47/69.6 | Internet and computer training; ensure interactions; provide conversation topics | Home/pilot study | Depression, tech use and usability/social support and isolation | PROMIS social isolation 6-i, social Provisions scale/no significant difference |
aThe number of participants in controlled studies is shown in parentheses (intervention; control); participants' age is indicated as mean, age range, or as reported in the study. bLS stands for Loneliness Scale.
Six interventions conducted qualitative studies, relying on direct or indirect (e.g., reports by staff) observation, questionnaires, and interviews (see Table
The other 19 interventions conducted quantitative studies, although only five were randomized controlled trials (RCTs) [
All but two interventions had measured participants’ conditions at baseline, indicating that being lonely or isolated was a requirement for inclusion but without reporting how this condition was determined [
Finally, we mention the lack of agreement on the effectiveness of video chat and social networks. Széman [
With respect to our first research question, we found that the
In answer to our second research question, we found that Internet access was fundamental to support long-distance interactions in all interventions, except for the telephone befriending service [
Technology used in interventions.
Intervention | Technology | Custom or off-the-shelf | Devices | Technology ownership/experience | Training or support |
---|---|---|---|---|---|
Ballantyne et al. [ | Social network (About my age) | Off-the-shelf | Computer | Nonproficiency required | Initial training sessions |
Blažun et al. [ | E-mail, Internet, and Skype | Off-the-shelf | Computer | Nonproficiency required | Initial training sessions |
Cattan et al. [ | Phone calls | Off-the-shelf | Telephone | N/A | N/A |
Cotten et al. [ | E-mail, Internet, and Facebook | Off-the-shelf | Computer | Did not report | Initial training sessions |
Dodge et al. [ | Video chat | Custom | Touch-screen computer | No previous use of PC (15%) | Visits for setup; no training |
Fokkema and Knipscheer [ | E-mail and Internet | Off-the-shelf | Computer | Nonproficiency required | Initial training sessions |
Garattini et al. [ | Virtual room, calls, messages, and broadcasts | Custom | Touch-screen computer with phone handset | No computer ownership (68%) | Visits for training and support |
Larsson et al. [ | E-mail, Internet, Facebook, and Skype | Off-the-shelf | Computer | Computer ownership and nonproficiency required | Visits for training and support; remote training |
Machesney et al. [ | Virtual companion (pet avatar) | Custom | Tablet | Did not report | Continuous visits; remote support |
Széman et al. [ | E-mail, Internet, Facebook, and Skype | Off-the-shelf | Computer | Nonproficiency required | Initial training sessions |
White et al. [ | E-mail and Internet | Off-the-shelf | Computer | Owned a PC (9%); no previous experience (60%) | Continuous visits; remote support |
Baez et al. [ | Virtual classroom, messages, and predefined messages | Custom | Tablet | Did not report | Initial training; remote support |
Czaja et al. [ | E-mail, Internet, virtual classroom, and messages | Custom | Computer | Participants had minimal computer or Internet use experience | Initial training; check visits; remote support |
Banbury et al. [ | Skype | Off-the-shelf | Tablet | Most had no previous video conference experience | Visit for setup |
Barbosa et al. [ | Messages (video, photos, audio, and predefined) | Custom | Tablet | Moderate (5), basic (3), or no experience (4) | Initial training; weekly visits for support |
Chiu and Wu [ | Line (messaging service) and YouTube | Off-the-shelf | Tablet | No computer learning experience (82%) | Training sessions (long period) |
Gutierrez et al. [ | Video chat, messages, and photos | Custom | Tablet | First-time as computer user required | Provided by family member |
Isaacson et al. [ | Virtual classroom, video chat, and photos | Custom | TV, remote, and webcam | Many not proficient with smart phones/computers | Visit for setup and training |
Jarvis et al. [ | WhatsApp (messaging service) | Off-the-shelf | Mobile phone (smart phone) | Used mobile to contact family and friends (55%), none had used WhatsApp | Initial training; weekly visits for support |
Jarvis et al. [ | WhatsApp (messaging service) | Off-the-shelf | Mobile phone (smart phone) | N/A | Initial training; weekly visits for support |
Morton et al. [ | E-mail, Internet, Facebook, and Skype | Off-the-shelf | Touch-screen computer | Required no current access to Internet | Continuous visits; remote support |
Myhre et al. [ | Off-the-shelf | Computer and tablet | No social network or minimal use required, tablet/computer ownership required | Initial training | |
Barbosa et al. [ | Video chat, photos, audio recording, and predefined messages | Custom | Tablet | All participants inexperienced with tech, save one | Initial training; weekly visits for support |
Pauly et al. [ | E-mail, Internet, social network, and messages | Off-the-shelf | Tablet | None or very little experience with portable electronic devices (67%) | Initial training; workshop during intervention |
Tomasino et al. [ | Virtual classroom, messages | Custom | Computer, tablet, mobile phones | Required Internet access and basic Internet skills | Remote support |
Off-the-shelf solutions were favored (
Computers, along with the mouse and keyboard as input devices, were preferred for supporting interventions (
While interventions clearly report on the technology and devices used, features and channels used for communication are less discussed. This information is useful to gain insight into participants’ preferences and adoption, and it is usually reported in terms of “most used” features. Some interventions report that comments and likes are preferred over predetermined messages [
In terms of understanding the human factors in the interventions and the relation between users and technology, we found that all interventions required older adults to use the technology on their own, although some studies reported that assistance was necessary for a long period [
With respect to the difficulties in interacting with technology, White et al. [
Tablet users reported feeling silly talking to a virtual pet, problems with audio, and delay in messages [
Usability, although not formally an outcome, was analyzed by some interventions. In a computer training course by Blažun et al. [
In relation to our third research question, out of the 25 interventions, 20 involved online groups (see Table
Social interactions and contacts.
Intervention | Online group or one-to-one | Contacts | Contact with research staff |
---|---|---|---|
Ballantyne et al. [ | Online group | Family and friends, new people | Weekly visits first, then fewer; phone calls at most 1 h/week |
Blažun et al. [ | Online group | Family and friends, new people | Training once a week; 4 h in Finland and 3 h in Slovenia |
Cattan et al. [ | One-to-one | Volunteers (predefined) | Variable number of weekly calls |
Cotten et al. [ | Online group | Family and friends, new people | Eight-week training (data from first 2 weeks) |
Dodge et al. [ | One-to-one | Predefined (trained interviewers) | Video chat 30–35 min/day; 5 days/week |
Fokkema and Knipscheer [ | Online group | Family and friends, new people, other participants, acquaintances | 5 × 2 h lessons; visits every 2–3 weeks |
Garattini et al. [ | Online group | Family and friends, other participants | 4 × 1 h visits; messages via app; weekly calls (extra calls for technical issues) |
Larsson et al. [ | Online group | Family and friends, new people, and other participants | Individual meeting offered weekly, group meeting every 2 weeks |
Machesney et al. [ | One-to-one | Predefined (trained helpers) | Visits and phone calls, available 24/7 |
Széman et al. [ | Online group | Family and friends, new people, and acquaintances | 1 × 1.5 h lesson; 1 h visits twice a week |
White et al. [ | Online group | Family and friends, new people | 3 × 2 h lessons, three 1 h lessons; trainer visits 2 h/week |
Baez et al. [ | Online group | Other participants, predefined (coach) | 1.5 h training model before baseline; support messages/calls |
Czaja et al. [ | Online group | Family and friends, other participants | Initial setup, 3x check visits, calls at week 1, months 3 and 9 |
Banbury et al. [ | Online group | Family and friends, predefined (facilitator), and other participants | Minimal training |
Barbosa et al. [ | One-to-one | Family and friends, acquaintances | Individual training at before deployment; weekly support visits |
Chiu and Wu [ | Online group | Family and friends | 90-min ICT training sessions weekly for 12 weeks |
Gutierrez et al. [ | Online group | Family and friends | Unaccounted frequency, support provided by family member |
Isaacson et al. [ | Online group | Family and friends, other participants | Visit for setup and training and after 4–5 weeks |
Jarvis et al. [ | Online group | Family and friends, predefined (facilitator) | 8x 90-min training session (during first 4 weeks), then weekly support visits |
Jarvis et al. [ | Online group | Family and friends, other participants | 2x 90-min training session/week over 15 days, then weekly support visits |
Morton et al. [ | Online group | Family and friends, other participants, and acquaintances | 3x 90-min training/week (month 1); session every 2 weeks, 1 h remote support alternate weeks (month 2), then 1 month of remote support and no visits |
Myhre et al. [ | Online group | Other participants | 3 × 2 h training session for a week |
Barbosa et al. [ | One-to-one | Family and friends | 1x individual training session, then weekly visits |
Pauly et al. [ | Online group | Family and friends | 2x training sessions before intervention, 3 h customized workshop during intervention |
Tomasino et al. [ | Online group | Other participants, predefined (coach) | Remote support throughout the study |
The majority of interventions (
Interestingly, Széman [
One-to-one interventions limited contacts to family, friends, and acquaintances [
Intergenerational relationships were also indicated as important. Three computer training interventions explicitly mention interactions with young people. Blažun et al. [
Interactions in person also occurred. Sometimes, participants met during the interventions and formed groups: a computer interest group [
In addition, visits were made in order to provide assistance or to make sure that systems were working properly (see Table
In the following, we analyze the findings from our research questions. In terms of challenges addressed by interventions, the strategies applied, and the intervention outcomes (RQ1), we observed that most interventions have dealt with the lack of social relationships and infrequent contacts by training participants in the use of computers and Internet. While results have been positive, and it is true that training participants or providing simple technology might solve the digital divide, such strategies do not guarantee access to contacts or frequent interactions. We argue that it is important to address the barriers directly, targeting challenges with technology that incorporates strategies by design. Interventions providing simple technology also ensured interaction as a strategy, and most of these interventions have reported positive outcomes. More studies taking this strategy would contribute evidence allowing for comparison with studies that train participants. Also, some studies have tried to improve conversations, by providing some contextual information and conversation topics. This area seems promising, especially in light of the surge of artificial intelligence and conversational agents. As fully automated conversational agents were successful on interventions for young adults with symptoms of depression and anxiety [
We also note that future studies should look into providing stronger evidence on the impact of the interventions conducted. Despite a majority of interventions reporting positive outcomes, and relying on quantitative methods, only five were RCTs. Studies disagree on the effectiveness of the technologies used (e.g., video chat and social networks), and some qualitative studies reports were obtained without standard measurement tools [
While we are advocating for more rigorous methodologies, we are not suggesting that qualitative outcomes should be abandoned in favor of quantitative ones. Qualitative studies are indeed useful to provide insight, especially on the motives behind behaviors. In this review, we have found studies reporting on the reasons for technologies not being adopted or reasons for people to interact more (or less) with others, uncovered through qualitative methods. In this sense, a methodological guide to designing studies in this area would greatly benefit researchers, especially those from the IT field, who might be less familiar with user studies with vulnerable subjects.
With respect to the technology used in interventions and how this technology was used by older adults (RQ2), we found that desktop/laptops make for a big part of the devices used to support interventions. Considering how commonplace long-distance interactions are nowadays and the availability of devices (e.g., mobile phones), we were surprised to find that studies based on desktop computers were so common, especially, since age-related limitations experienced by older adult computer users [
We also found less solutions designed specifically for older adults than studies relying on off-the-shelf technologies. This might be due to the higher investment that designing tailored technologies requires. Nonetheless, previous research shows that tailored tools could increase adoption [
The prevailing technologies were e-mail and general Internet use for interaction (e.g., discussions in forums), closely followed by social networks and video chat. We must note, however, that while we have identified some studies using more recent technologies such as messaging services (e.g., WhatsApp) that allow for picture sharing and reaching relatives more conveniently, we still observe a disconnect between the latest technologies available and those used to conduct formal studies.
Many interventions enabled a combination of features and channels for interaction; however, few reported on how and how frequently these were used. Since such reports are scant, we cannot assess technology adoption or effectiveness. Therefore, we recommend future interventions to add formal reports on usability (e.g., the System Usability Scale [
Finally, with respect to the social interactions enabled (RQ3), we found that most interventions enabled interaction with online groups, rather than with one person put in place specifically for the intervention. Family and friends were the contact group reported by a majority of the interventions, some highlighting intergenerational relationships as particularly important for older adults. Nonetheless, here, we also lack quantitative information on the frequency of contact. Since all interventions with online groups included at least two different groups of people (e.g., family and friends and other participants), we cannot tell whether older adults prefer to contact certain groups nor assess the impact the type of relationship has on the effectiveness of interventions. Friendship relationships, for instance, have been associated with stronger effects for subjective wellbeing [
The need for quantitative information also applies to the channels used to interact with people from particular groups. For example, Széman [
Furthermore, despite assessing the effect of long-distance interactions, many studies reported interactions in person during the intervention (e.g., with other participants, with researchers). If interactions were frequent, the effect on intervention outcomes should be considered.
For interventions, technology had the fundamental role of enabling long-distance interactions and was used for support in different ways. By facilitating more channels for interaction and providing access to larger audiences, it allowed participants to expand social networks, strengthen existing ties, providing social support, or build community rapport. However, since existing interventions are few, they tell us about the feasibility of using technology for long-distance interactions, but it is still unclear how technology is actually used, what limitations and opportunities exist, and how these affect the success of the intervention.
Therefore, we highlight some recommendations for researchers approaching this field of study. First, on the study methods, it is important to (1) design studies as RCTs, (2) leverage standard instruments for measuring loneliness and social isolation, and (3) consider the potential impact of continued (and in person) contact with participants on measurements. This may seem obvious, but we found few studies with these characteristics. We also recommend to report and discuss separately the results for each interaction channel and by the type of relationship (e.g., with friends, children, and grandchildren), since without this information, it is hard to infer what worked. Second, in terms of challenges, open opportunities lie in studying how technologies can facilitate and improve conversation (e.g., by presenting shared interests as topics), as opposed to enabling them. Finally, a vast majority of current research has focused on training for using a specific technology. To date, little attention has been paid to (1) designing interventions that enable or encourage usage of technology in specific ways (e.g., organizing and encouraging access to chat rooms with specific topics) and on (2) using persuasive technologies that introduce motivational elements and help users initiate and sustain conversations on shared interests. We feel that addressing these gaps in current research can lead to a better understanding of the role technology can play in tackling loneliness, helping to alleviate one of the modern ailments of our society.
Part of this work has been included and published as part of the doctoral dissertation entitled “The Connective Power of Reminiscence: Designing a Reminiscence-based Tool to Increase Social Interactions in Residential Care” [
The authors declare that there are no conflicts of interest regarding the publication of this paper.
This work was supported by the Russian Science Foundation (project no. 19-18-00282). The authors would like to express their gratitude to the authors of the papers listed in this review who replied to their inquiry for further details. Their help has made this review work more thorough.