Literature search
Electronic searches identified 529 citations, resulting in 485 unique citations to be screened for inclusion following removal of duplicates (see Fig. 1). The titles and abstracts were assessed for their relevance to the review based on the inclusion criteria (Stage 1 screening), resulting in 46 citations being retained. The full texts of all these citations were obtained and after applying the inclusion criteria (Stage 2 selection), 30 citations were excluded; 12 did not provide data relevant to categorisation of loneliness and/or social isolation intervention(s), 15 were not reviews and three did not have a primary or secondary objective of reducing loneliness and/or social isolation. An additional 17 citations were identified through backward citation chaining and these citations were also included. As such, 33 citations were included in the scoping review (see Fig. 1). Characteristics of the included reviews are shown as a structured table and as a narrative summary in Additional file 2: Table S2.
Characteristics of reviews
There is increasing interest and research in the area of loneliness and social isolation among the older population. The first review appeared in 1984 and following that, there were three more reviews up until the year 2003. Subsequently, there were more frequent publications of literature on loneliness and/or social isolation and at least one review was published consecutively every year from 2010 onwards. This information is represented in a diagrammatic form in Fig. 2.
Type of reviews
Review papers were published between 1984 and 2017 and of these, systematic reviews were the most common type of reviews obtained [9, 10, 26, 31,32,33,34,35,36,37,38,39,40,41], followed by literature reviews [6, 42,43,44,45,46,47], evidence reviews [18, 48,49,50], narrative reviews [25, 51, 52], and other types of review including critical [53], empirical [54], rapid [55] and integrative review [11]. This information is represented in a diagrammatic form in Fig. 3.
Of those reviews which employed a systematic means of selecting eligible primary research (n = 14), ten papers included only studies published in English, two review papers included studies published in any language, one review included studies published in English and Italian [34], and 1 included studies published in English, French, Italian and Spanish [41].
Concept of loneliness and social isolation
In terms of the consideration of the concepts of loneliness and social isolation, most reviews (28/33) could be assigned to one of three categories: 1) reviews that explicitly focused on interventions to reduce social isolation (n = 4) e.g. Chen and Schulz [37], Findlay [33], Oliver, et al. [47] and Wilson and Cordier [52]; 2) reviews that explicitly focused on interventions to alleviate loneliness (n = 11), e.g. McWhirter [6] and Masi, et al. [31] and Cohen-Mansfield and Perach [53]; and 3) reviews that included papers with interventions for both loneliness and social isolation (n = 13) e.g. Poscia, et al. [34] and Cattan, et al. [10]. The remaining five reviews focused on loneliness and other outcomes of interests such as anxiety and depression (n = 3); or other related concepts such as social participation [56], and social connectedness [35]. While there is a distinction between loneliness and social isolation, there was not any obvious differences in reviews that focused on loneliness or social isolation in terms of the review type, where the research was conducted, and how the findings were reported.
Loneliness/social isolation was not always reported as the primary outcome and was sometimes reported alongside other health outcomes as seen in three reviews [36, 38, 40]. A review by Choi, et al. [40] examined the effectiveness of computer and internet training on reducing loneliness and depression in older adults. Elias, et al. [38] evaluated the effectiveness of group reminiscence therapy for loneliness, anxiety and depression in older adults. In a review by Franck, et al. [36], interventions were reviewed if they addressed social isolation, loneliness, or the combination of depression with social isolation or loneliness. In a systematic review by Morris, et al. [35], the effectiveness of smart technologies was examined in improving or maintaining social connectedness.
Population characteristics
The majority of the reviews (n = 24) focused solely on the older population [9,10,11, 18, 25, 26, 32,33,34,35,36,37,38,39,40,41, 46, 48,49,50,51, 53, 55, 56] but the age range used to define this population varied [32, 35,36,37,38, 53], or was not specified at all [9,10,11, 26, 33, 46, 51, 56]. For example, a systematic review by Morris, et al. [35] targeted older people who live at home and included participants that were aged ≥45 years, whereas Cohen-Mansfield and Perach [53] and Chen and Schulz [37] targeted individuals aged ≥55 years; and Chipps, et al. [32], Franck, et al. [36], and Elias, et al. [38] targeted individuals aged ≥60 years. Where age was not specified, review authors used the term ‘older people’ or its synonyms, e.g. older adults [40] and seniors [39, 56], to describe the target population. It was stated in two of these reviews that the definition for the older person was defined by the criteria used in the studies included in the review [26, 56].
Some reviews focused on specific subgroups of the older population which research has identified to be more prone to loneliness and social isolation. For example, six reviews focused only on older people residing within the community [6, 10, 39, 42, 48, 52], whereas three focused only on older people living in institutionalised settings e.g. care or nursing homes [36, 38, 47]. The majority of reviews (21 in total) included populations of both community-dwelling individuals and those living in long-term care [9, 11, 18, 25, 31,32,33,34,35, 37, 40, 41, 44,45,46, 49,50,51, 53, 55, 56]. Residential status was not reported in three reviews [26, 43, 54]. This population characteristic is represented diagrammatically in Fig. 4.
Only two of the 33 reviews [41, 52] included gender as an inclusion criteria, and these two papers focused specifically on interventions targeted at the male population only, including Men’s Sheds [52] and gendered interventions for older men [41]. Men’s Sheds are community-based organisations that provide a space for older men to participate in craftwork and engage in social interaction [52]. Review authors often reported that the gender distribution of participants in primary research involving loneliness/social isolation was heavily skewed towards the female population [31, 33, 34, 37, 39]. The subsets of the female population reported in the reviews includes: isolated older women, women at risk of suicide, senior women on the housing waiting list [33]; community-living, chronically ill women [31]; women with primary breast cancer, community-dwelling, single women [34]; and community-dwelling low-income women with low perceived social support [39].
Countries in which interventions were delivered
The countries in which interventions were delivered was not reported in some of the review papers (n = 16). Of the papers that did report this (n = 17), USA was the most reported (n = 14), followed by Netherlands (n = 13), Canada (n = 10), UK (n = 9), Australia (n = 8), Sweden (n = 8), Finland (n = 5), Taiwan (n = 5), Israel (n = 4), Norway (n = 4), Germany (n = 4), Japan (n = 3), China (n = 2), Hong Kong (n-2), Denmark (n = 2), Italy (n = 2), New Zealand (n = 2), South Africa (n = 1), Austria (n = 1), Slovenia (n = 1) and Iran (n = 1).
Categorisation of interventions
There was a broad range of terms that review authors used to describe the characteristics of interventions, such as: format [31], delivery mode [9, 31, 34], goal [42, 44, 46], type [9, 31, 34, 53], focus [53], and nature [46], and often the same terms had different meanings. Some authors used two or more categorisation systems as seen in the reviews by Dickens, et al. [9] and Poscia, et al. [34], where interventions were categorised by both their ‘delivery mode’ and ‘type’. Alternatively Grenade and Boldy [46] categorised interventions by their ‘nature’ and ‘goal’, and Cohen-Mansfield and Perach [53] categorised interventions based on their ‘focus’ and ‘type’. Masi, et al. [31] categorised interventions based on their ‘type, format and mode’. It was common (n = 20) for review authors to categorise interventions on the basis of whether they were delivered via a ‘group’ or ‘one-to-one’ [9, 10, 18, 26, 31, 33, 34, 36, 38, 39, 42, 44,45,46, 49,50,51, 53, 55, 56]. In a review by Raymond, et al. [56], social participation interventions were delivered in an individual or group context. Elias, et al. [38] explored the effectiveness of group reminiscence therapy in alleviating loneliness whereas the Medical Advisory Secretariat [39] evaluated in-person group-based interventions in alleviating loneliness and social isolation among community-dwelling care seniors. The term ‘mode’ was used frequently within review papers but often with inconsistent meanings. By way of illustration, Poscia, et al. [34] and Dickens, et al. [9] referred to the categorisation of interventions via group or one-to-one delivery as ‘mode’, and classified interventions as individual, group or mixed (both individual and group). In contrast, delivery ‘mode’ in Masi, et al.’s [31] review referred to ‘technology’ or ‘non-technology’ based interventions, and ‘format’ was used to describe whether the intervention was implemented on a one-to-one basis or as a group (if more than one person participated in the intervention at the same time or if the intervention involved asynchronous interactions such as internet-based chat room exchanges).
Some review authors categorised interventions by their type (n = 4) [9, 31, 34, 53], and the descriptions for this category also varied. In a review by Dickens, et al. [9], interventions categorised by their ‘type’ were described as: ‘offering activities’ (e.g. social or physical programmes), ‘support’ (discussion, counselling, therapy or education), ‘internet training’, ‘home visiting’ or ‘service provision’. In another review, intervention type was described as: 1) social skills training if it focused on improving participants’ interpersonal communication skills, 2) enhancing social support if the intervention offered regular contacts, care or companionship, 3) social access if the intervention increased opportunities for participants to engage in social interaction (e.g. online chat room or social activities), and 4) social cognitive training if the intervention focused on changing participants’ social cognition [31].
Similar to the categories used in the review by Masi, et al. [31], Poscia, et al. [34] categorised interventions by their type, further describing the category as offering: [1] social support (e.g. discussion, counselling, therapy or education), 2) social activities, in form of social programmes, 3) Physical activity (fitness programme or recreational activity), 4) technology (e.g. companion robot, telephone befriending or internet use), 5) singing sessions, and 6) horticultural therapy. By contrast, when Cohen-Mansfield and Perach [53] categorised interventions by their ‘type’ this referred to whether interventions were delivered in a ‘group’ or ‘one-to-one’.
Three reviews categorised interventions by their ‘goal’ [42, 44, 46]. In two of these reviews [42, 44] the same constructs were used to define goals and these were: 1) to facilitate social bonding e.g. via cognitive behaviour therapy or social skills training, 2) to enhance coping with loneliness e.g. through support groups, and 3) to prevent loneliness from occurring e.g. through community awareness and educational programs. In the third review [46], the authors implicitly addressed these three constructs but used different terminology, i.e. to enhance people’s social networks, and promote personal efficacy and behaviour modification, and/or skills development. A similar categorisation system was used by Cacioppo, et al. [45], but these review authors labelled this category as ‘models of loneliness interventions’ rather than ‘goal’, and included interventions aimed to: 1) provide social support, 2) increase opportunities for social interaction and 3) teach lonely people to master social skills.
A total of six reviews focused on technology-based interventions to improve communication and social connection among older people [32, 35, 37, 40, 47, 48]. An evidence review by Age UK [48] reviewed the use of modern (e.g. internet) and assistive technology (e.g. telecare or telehealth) in maintaining and establishing social contact. Chen and Schulz [37] reviewed the effects of communication programs such as telephone befriending, computer and internet, and high-technology apps such as virtual pet companions in reducing loneliness and social isolation in the elderly. The effectiveness of e-interventions which can be described as online activities e.g. computer or internet training and usage; interpersonal communication e.g. Skype; and internet-operated therapeutic software e.g. Nintendo Wii entertainment system and videogames, were synthesised and assessed for decreasing social isolation and loneliness among older people living in community/residential care [32]. One systematic review evaluated the effectiveness of smart technologies [35], which can be described as internet-based support groups and computer use and training, whereas the potential of videophone technology in improving communication between residents and family members was reviewed by Oliver, et al. [47]. In another review, computer and internet training among lonely and depressed older adults were examined [40].
The rationale for the categorisation of interventions was reported in the majority of reviews (n = 21). It was stated in an integrative review by Gardiner, et al. [11] that interventions were categorised based on their purpose, intended outcomes and mechanisms by which they targeted loneliness and social isolation. Gardiner, et al. [11] highlighted the importance of this categorisation given the growing diversity in intervention types, and considered rigorous and transparent categorisation to be a necessary pre-requisite for identifying which elements of interventions influence their effectiveness. Their thematic synthesis identified six categories which included: social facilitation interventions, psychological therapies, health and social care provision, animal interventions, befriending interventions, and leisure/skills development. In a narrative synthesis by Jopling [25], interventions were grouped in accordance to addressing three key challenges: 1) reaching lonely individuals, 2) understanding the nature of an individual’s loneliness and developing a personalised response, and 3) supporting lonely individuals to access appropriate services.
Other reviews [36,37,38,39,40,41, 47,48,49,50,51, 55] did not report a rationale for the categorisation of interventions (as seen in Additional file 2: Table S2). Some review authors justified their categories on the basis that they had been used in previous reviews, e.g. two reviews [10, 33] replicated the categorisation used in a previous study by Cattan and White where intervention studies were divided into four categories based on the programme or method type, i.e. group activity; one-to-one intervention; service delivery; and whole community approach [26]. Likewise, McWhirter [6] used similar categories as Rook and Peplau [57], such as cognitive-behavioural therapy, social skills training, and the development of social support networks; Andersson [42] categorised interventions based on the typology of social network interventions by Biegel, et al. [58] (either clinical treatment, family caretaker enhancement, case management, neighbourhood helping, volunteering linking, mutual aid/self-help, and community empowerment); and Masi, et al. [31] categorised the intervention type (i.e. providing social access, social cognitive training, social skills training or social support) based on similar constructs used in the reviews by Rook [44], McWhirter [6], Cattan and White [26], Findlay [33], Cattan, et al. [10] and Perese and Wolf [43].