Reducing loneliness among migrant and ethnic minority people: a participatory evidence synthesis

Initial database searches (approach 1)

Initial searches of electronic databases sought studies in which the target population was identified as either (1) migrant and/or ethnic minority people (however defined and labelled by authors), or (2) a client group, population or neighbourhood diverse in terms of ethnicity/migrant status, and for which the study included a focus on the outcomes of interest. Early searches confirmed the importance of widening core terms/labels (social isolation, isolation, loneliness) to include (sense of) belonging, social connectedness, social networks, social ties and social relationships. It also revealed a need to consider studies that included outcomes that are a potential part of a causal pathway to social isolation and loneliness. A comprehensive list of search terms was developed iteratively from exploratory searches, medical subject heading (MeSH) terms and harvesting keywords from relevant reviews. As anticipated, indexing of social isolation and loneliness outcomes was variable, requiring identification of appropriate free-text terms. Searches of three key electronic databases (MEDLINE, Applied Social Sciences Index and Abstracts, and Social Science Citation Index via Web of Science) were undertaken in October 2017, using free-text and MeSH terms (see Appendix 1 for examples). No date restrictions were applied. A UK filter was initially tried, but this was not successful across the databases; therefore, we opted to retrieve international literature.

Review of prior reviews (approach 2)

We conducted a review of reviews to engage with the wider literature (theoretical and empirical) around social isolation and loneliness (i.e. not specific to migrant/ethnic minority people). All types of review, including evidence from UK and international studies, were included, not just those that were intervention related. We adopted a two-stage approach, with an initial batch of reviews being read and extracted in detail using a Microsoft Excel^® (Microsoft Corporation, Redmond, WA, USA) template, and subsequent reviews being screened more quickly so that only material that added to the emerging theoretical picture was extracted. All of the prior reviews were examined carefully for identification of relevant intervention-/initiative-related studies. The review of reviews surfaced definitions and theories around loneliness, social isolation and related concepts. It also identified prior intervention typologies, informing our own thinking on the utility of categorisations, which was explored with the Project Advisory Group (PAG). Identification of intervention terms/labels revealed that some were too generic to be usefully employed in further searches, for example ‘support group’. We therefore decided, with agreement from the PAG, to focus on terms implicitly about isolation or loneliness, for example ‘friendship group’. Further reviews of potential relevance were identified during screening and extraction for the other review components, and via ongoing literature alerts, revealing this as a very active area of research. These were examined to (1) ensure that no relevant empirical intervention studies were overlooked and (2) identify useful theory to inform the review approach.

Additional search techniques (approaches 3, 4, 5 and 6)

We undertook targeted electronic database searches in October 2018 (approach 3) using a list of interventional approaches compiled through the aforementioned stages of searching and sifting, and suggestions from the PAG and CPs (e.g. befriending/companionship, cognitive–behavioural therapy/counselling, friendship groups, neighbourhood cohesion, shared spaces/housing design), and a list of loneliness measurement scale terms.

Complementary searches were conducted by forward and backward citation-searching (approach 4) using Publish or Perish information management systems (Publish or Perish Inc., Houston, TX, USA) and the reference lists of included intervention studies. Typically, most included studies cited between eight and 20 papers, although older or influential studies cited considerably more than this. The yield from citation-searching compares favourably with screening from electronic subject searches, and search methods using forward and backward citation-tracking are particularly valuable in finding documents to develop and test provisional theories. 48

Members of the PAG also made recommendations (approach 5), particularly in terms of theoretical material and prior reviews. Throughout the project period, journal alerts (approach 6) were used to capture newly published material up to the end of August 2019.

Grey literature identification (approach 7)

Grey literature was included to identify UK-based interventions/initiatives of relevance. We identified grey literature by Google searches (Google Inc., Mountain View, CA, USA) using key terms (both generic and intervention-specific, as harvested from earlier searches); searching websites of relevant policy, practice, research and advocacy organisations and contacting their research and evaluation departments by e-mail; contacting subject experts and using subject-relevant e-mail distribution groups; and searching OpenGrey. Grey literature searching was completed in July 2018.

The search for empirical literature was recorded for audit trail purposes. The search process was undertaken according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines. 49

Recruitment and participants

Consultation panel participation was based on interest in the topic, an ability to relate the topic to real-world experiences and a capacity to consider the influence of broader social processes. It was not a requirement that participants were lonely or isolated. The purpose of the CPs was not to elicit revealing or distressing personal stories, but to offer insight into processes through which loneliness and social isolation arise and can be addressed. We decided to convene three CPs based around broad age/life-stage groups: young adults/students, working-age people and older people. Potential participants were identified in partnership with community-based organisations and existing black and minority ethnic (BME) and migration-oriented networks. After interested respondents had been approached and given an opportunity to reflect, researchers contacted them by telephone to discuss the project in more detail. Members were selected to ensure broad representation across social characteristics (e.g. gender, ethnicity), as well as their active contribution. We recruited between six and 11 individuals to each CP (two in Sheffield and one in Leicester). Overall, 34 participants contributed. Two interested participants, from a minority group that was under-represented in the CP workshops, were included in one-to-one telephone and face-to-face discussions as they were unable to attend the workshops. Participants were members of the public and people working in community-based roles with good knowledge of local communities. All participants self-identified as a migrant and/or having an ethnic minority identity. The groups included recent arrivals and more established migrants, as well as UK-born individuals. One panel included people who had been forced to migrate. Individuals who did not speak English fluently were not excluded. One participant was provided an interpreter to assist during workshops. Our approach to the CPs did not make any prior assumptions about differences in experiences between migrant and UK-born ethnic minority people. Exercises encouraged participants to examine the commonalities and divergences within and between groups of people ascribed particular ethnic and migration labels.

The management and ethos of the consultation panels

Workshops were held in local venues familiar to many of the participants. Timings were arranged in consultation with members.

The approach was based on participatory principles set out by INVOLVE. 51 Our approach was informed by Group Model Building methods,52 which emphasise the importance of adopting an iterative process of dialogue, building, testing and improvement, and the need to attend carefully to group dynamics. Early ground rules were collaboratively developed during introductory sessions to ensure that everyone had the opportunity to create an enabling ethos. Important principles suggested by CP members were respectful dialogue, listening carefully and making sure that information shared among the group was not divulged beyond the group.

Participants were offered £50 shopping vouchers to recognise their contributions to the half-day workshops. All travel expenses and any care costs were paid.

Structure and content of consultation panel sessions

One introductory session and two half-day workshops were organised for each group. A final co-creation session with CP members from all the three groups was convened to generate visual outputs.

Consultation panel contribution

Consultation panel workshops contributed importantly to the review in the following ways:

• critique and expansion of the concepts of social isolation and loneliness, based on lived experience

• generation of theory, through panel member narratives, on how social isolation and loneliness operate across different levels of the system

• shaping the inclusion and typing of interventions and initiatives

• identification of upstream, mid-range and individual factors that can influence the effectiveness of interventions

• commentary on, and refinement of, draft models of proximate and distal causes of social isolation and loneliness

• identification of why interventions/initiatives are more or less accessible, and work more or less well, for different people

• identification of some gaps and uncertainties in the team’s analysis that warranted follow-up, with such insights informing follow-on database searches.

At each stage, our analysis drew across the insights from the three CPs to highlight elements of the constructed theory that were: supported, refuted, or refined/extended by CP testimony, as well as new system characteristics not previously identified.

Concepts and definitions

Loneliness can be considered a ‘chaotic concept’; variously defined and measured, debated and contested. From a public health perspective, loneliness is seen as a problem because it has been associated with a range of negative physical and mental health outcomes. Growing concerns about levels of loneliness, particularly among the elderly, might also be a seen as a moral panic, as politicians, civil society and the media invoke images of the breakdown of the social fabric of families and communities.

It is now commonly accepted that social isolation and loneliness are not the same; the former relates to the objective state in which an individual has little contact with family members, friends or neighbours,53 and the latter relates to a subjective assessment that one’s social relationships are inadequate. This distinction is useful because it draws our attention to the quality of social relationships and interactions, and suggests that a person who, at face value, is not isolated from other people can, nevertheless, be lonely. At the same time, the distinction also suggests that people who have little social contact may not necessarily feel lonely. Although true, research does suggest a strong correlation between the number and intensity of social ties and the risk of loneliness. 54 Individuals who live solitary lives without feeling alone seem to be the exception. Understanding loneliness as resulting from the interplay of a deficit in social connection and an individual’s cognitive assessment of the adequacy of relationships seems likely to be more fruitful. 55

It has been suggested that loneliness can arise because an individual perceives any or all of the quantity, quality or modality of their relationships to be lacking. 56 There is also the issue of which relationships, when perceived to be lacking, can generate loneliness. Some authors have drawn a distinction between ‘intimate’ and ‘social’ loneliness; the former relates to perceived inadequacy in close, affirmative, confiding ties (typically family members and close friends), whereas the latter relates to a perceived lack of relationships that provide camaraderie. Questions are raised about whether different types of relationships are functionally distinct and necessary for emotional well-being, providing an individual with different forms of social sustenance, or whether they are, instead, interchangeable. Common definitions and measures of loneliness frequently merge both intimate and social dimensions, such as Rook’s:4

[A]n enduring condition of emotional distress that arises when a person feels estranged from, misunderstood or rejected by others and/or lacks appropriate social partners for desired activities, particularly activities that provide a sense of social integration and opportunities for emotional intimacy.

Rook4

Although Rook’s4 definition suggests an ‘enduring’ condition, the need to recognise variation in the temporal and spatial characteristics of loneliness has also been noted. Loneliness may be a temporary state, often linked to life stages or transitions or a change in circumstances, or it may be long lasting. Loneliness may also be situationally specific, experienced only in a particular sphere or space of life, such as the home, the workplace or at school. The intensity of loneliness also varies. For some, it is an extreme state of emotional distress, whereas for others it is less intense. 56 These elements of chronicity versus transience, generalised versus situation specific, and variations in intensity can co-occur in any combination. Some measures of loneliness use scales that could, in theory, enable assessment of the degree of intensity of loneliness, although this approach has been rare to date. Longitudinal data on the fleeting versus enduring nature of loneliness are also rare. Some research focuses on experiences of loneliness in particular settings, often framed in terms of ‘sense of belonging’, a concept that has particular resonance for our population groups of interest, as discussed more in Sense of belonging and racism. To date, there is little evidence of the implications of these different forms of loneliness for physical and mental health across the lifespan.

Social support

The distinction between the quality and the quantity of social relationships found in theories of loneliness and social isolation is mirrored in the literature concerned with the conceptualisation and measurement of ‘social support’. For instance, Gottleib and Bergen62 note the importance of distinguishing between the structure of a person’s social network on the one hand, and the resources that arise from such social ties on the other, eschewing the suggestion that ‘people’s social ties are unconditionally supportive’. 62 Theoretical work in this field has highlighted the importance of examining both sources of support and the types of support that are available and/or provided.

Gottleib and Bergen62 define social support as ‘social resources acquired from non-professionals in the context of both formal support groups and informal helping relationships’, whereas others also regard professionals as potential sources of support. 63

Work in the 1980s sought to identify types of social support. Barrera64 identified five types of social support: emotional, instrumental, companionship, informational and esteem support; House and Kabarl65 opted for four types of social support: emotional, tangible, informational and appraisal. Barrera64 also argued that a distinction should be drawn between perceived support and that which is actually mobilised and received, and empirical research62 has tended to show that it is people’s sense of support that provides a buffering effect in stressful or challenging situations. Clearly, the concept of perceiving a lack of available social relationships that can provide emotional and affirmational support and companionship is akin to the definitions of loneliness commonly employed.

In addition, social support literature draws attention to the qualitative adequacy of the support offered, that is the manner and associated meaning that is conveyed through the provision of support. An offer or provision of social resources that is made in an unsuitable way can be unsupportive, or even damaging to the recipient, even if the intention was positive. 62 Spontaneous receipt of social resources from one’s social network may also have a more positive impact than that which has to be requested. 62

These theoretical contributions enhance the loneliness literature by drawing clearer attention to so-called ‘negative social support’, that is interactions and social ties that are experienced as stressful, that are unsupportive and that are detrimental to the individual’s mental well-being. In this literature, ‘negative social support’ has most commonly been identified as emanating from familial relationships, particularly spousal and parent–children ties, that is relationships that are generally expected to provide emotional, affirmational and tangible support. 66

A further emphasis in the social support literature, that is less evident in the work on loneliness, is the reciprocal nature of many supportive social relationships: ‘social support is not a commodity that resides in the provider and passes to the recipient, but . . . is an expression of the mutuality and affection characteristic of the relationship between the parties’. 62 ‘Peer support’ has received particular attention as a route via which effective emotional and information support can be delivered, as it is predicated on the principle of mutuality. 62,67 Non-reciprocal relationships will be perceived as supportive (rather than exploitative or degrading, depending on whether you are the giver or receiver) only if the unidirectional flow meets the expectations of both parties. The expectations people hold about relationships, shaped by individual attributes and collective values, are fundamental to their perceptions of social support. These conceptual contributions reveal the way in which social ties and flows of social resources are mutually constituting. A social relationship (which carries a particular meaning) presents an opportunity for the exchange of resources, the materialised nature and volume of which, in turn, serve to define and give meaning to the relationship. These insights have particular relevance to our focus on ethnic minority and migrant populations. Transition to an unfamiliar social context (whether due to international migration or movement into white-dominated institutions and spaces) can mean both the loss of supportive social ties and exposure to unaccustomed norms of relationship-building. 63,68 Furthermore, formal social support services may be deficient for migrants and minorities owing to a lack of cultural competence. 69

As well as providing useful insights into the meaning and consequences of interpersonal social exchanges, the social support literature usefully draws attention to the potential impact of ‘macrosocial’ variables, such as housing patterns, transport and job conditions, as well as status hierarchies (linked to gender, ethnicity, socioeconomic position), on people’s levels of perceived and actualised social support. 70 Contextual stressors such as poverty and being in an unfamiliar context can mean that social ties become overwhelmed or inadequate, and interactions more negative, creating the need for bolstering of informal support networks. 68

Social networks, social integration and acculturation

As noted above, social support researchers have argued that the structural properties of people’s social ties should be distinguished from their perceived or actual functions. Gottleib and Bergen62 suggest that structural properties of an individual’s social network (number of ties, density and interconnections) are measures of ‘social integration’, that is ‘the extent to which the person is enveloped in the social fabric.’62 Therefore, this notion of social integration is very similar to the concept of ‘social connectedness’ and, in essence, is the opposite of social isolation.

In addition, however, ‘social integration’ has been conceptualised more broadly, and variously, in relation to migrants (and, by extension, ethnic minority individuals). Here the interest has been on migrant social connections across various spheres of life. Ager and Strang’s71 widely cited model identifies 10 inter-related domains of integration and highlights ‘the fundamental role that social connection is seen to have played in driving the process of integration at a local level’. 71 However, although this and other formulations characterise integration as a process involving cognitive and behavioural adaptation on the part of both newcomers and established citizens, a so-called ‘two-way street’, leading to immigrants becoming accepted members of society,72 competing conceptualisations are also highly visible.

Linked to claims of self-segregating minorities who ascribe to values and norms that are incompatible with the host society (portrayed as homogeneous white British), and fears around strong intraethnic ties, ‘integration’ in policy circles has increasingly been understood as a process through which people of migrant and/or ethnic minority identity should adapt to the (imagined) ‘British’ way of life. 73 Such ‘assimilationist’ policy approaches have been critiqued from diverse angles, including their potential to undermine positive social ties within and across ethnic groups. 74

Parallel debates have occurred in relation to the entry of ethnic minority and/or migrant individuals into specific white majority-dominated social spaces and institutions. Some theorists advocate a process of adaptation to the new social way of life, through which newcomers replace their home culture with that of the new social setting. Others critique this approach as ‘cultural suicide’, arguing that successful outcomes require both acquisition of new cultural repertoires alongside the conservation of a sense of self and cultural tradition. 75

Theoretical work around the notion of ‘acculturation’ is also useful here. Berry76 identified the way in which people arriving in a new cultural context are faced with the need to accommodate both their heritage culture and that of the new society (or setting) in which they find themselves. This work drew attention to the varied patterns of association that migrants may establish with both people recognised as members of their ‘in group’ and those who are ‘out group’ members. Empirical work drawing on these ideas has tended to find that newcomers who manage to maintain a positive sense of their own ethnic identity and supportive ties to in-group members, while at the same time also forging positive ties to out-group members, have a lower risk of poor emotional health than those who have ties to either group alone, or no supportive ties at all. 77,78

These contributions are significant for the current project because they highlight the importance of considering the ethnocultural identity of individuals within social relationships and networks, and the implications of such identities for whether or not, and how, relationships meet individual social needs and expectations. Our CP members gave many examples of seeking out, and finding comfort in, relationships with individuals who shared their ethnic (or national) identity; ‘a little safety cocoon with yourself’, as one student CP member called it. At the same time, opportunities to form ties, and to interact informally, across difference were welcomed.

Sense of belonging and racism

Rooted in Durkheim’s79 early work on the social causation of suicide, Spady80,81 developed influential ideas around ‘belonging’, focused on a concern to reduce under-represented group dropout from US higher educational institutions. He identified the risk factors for dropout as ‘a lack of consistent, intimate interactions with others, holding values and orientations that are dissimilar from those of the general social collectivity, and lacking a sense of compatibility with the immediate social system’. 80 A large body of subsequent US work has explored the nature and determinants of ‘sense of belonging’ within predominantly white institutions. Strayhorn82 identified a sense of belonging as rooted in feeling accepted, valued, respected and cared for by peers and the wider institution. Baumeister and Leary83 have argued that the need to belong is a powerful, fundamental and pervasive motivation among human beings, and extends within and beyond particular institutional contexts. Hagerty et al. 84 identify belonging as the experience of personal involvement in a system or environment, together with the experience of being valued, needed and accepted, and ‘fit’ (perceiving that your characteristics articulate with or complement the system or environment).

These insights are particularly relevant to our focus on loneliness among migrant and ethnic minority people, whose sense of ‘mattering’, ‘being valued’ and ‘fitting in’ are frequently undermined. A large body of research is concerned with notions of belonging in a more general sense, to communities and to society at large, and how such belonging is constructed and negotiated by migrants and minorities in the face of exclusionary discourses and processes of ‘othering’. 74,85,86 This work also explores how sense of belonging may interplay with transnationalism. It is important to be alert to the ways in which the maintenance and disruption of transnational social networks may have implications for loneliness among migrants and minorities. Important also is the way in which transnational lives may lead to majority discourse that questions the ‘commitment’ of migrant and minority individuals to belonging where they reside and policies that undermine social connections. 74,87 Older CP members talked vividly about their loss of social connections outside the UK:

Because when we are older we yearn for things we had when we were younger, some of which you will never get back. That can become problematic in the mind . . . Food, connections with individuals who are actually not there anymore. The music, the conversations, the jokes, the understanding of what it is you left back home. Which isn’t there, by the way.

CP older member

The concept of self-worth, although present in the general literature on loneliness,88 is particularly foregrounded in the sense-of-belonging literature, and highly relevant to our focus on migrant and minority groups, whose identities are commonly stigmatised. This suggests the importance of examining the interplay between discrimination, stigma, shame and loneliness. Importantly, racial discrimination can usefully be understood as operating at multiple, interlocking levels. At a structural level, processes of racism are entrenched in laws, policies and practices, and are reproduced in societal institutions and organisations. At a cultural level, the circulation of images, language and symbols in multiple societal arenas – everyday conversation, media, the arts and policy discourse – serves to perpetuate, and normalise, the ideology of inferiority and difference. 89 These insights are important because, in addition to racist interpersonal interactions that might be experienced as outright verbal or physical abuse, or more subtle microaggressions, sense of belonging can be undermined by the material and symbolic properties of ‘white spaces’. 90

Intercultural encounters

A final related body of work seeks to understand patterns of interaction across ethnocultural difference, and how more positive social interactions, and reduced prejudice, can be achieved. Early psychological contributions to this area were framed in terms of ‘contact theory’,91 whereas more recent work has examined the nature of, and opportunities for, ‘meaningful encounters’ that can have a transformatory effect on attitudes and behaviours. 92,93 The importance of bringing this work into view for the present project was supported by our CP members, who underscored the relevance of both positive interethnic ties and negative interactions across ethnic difference, for sense of belonging among migrant and ethnic minority people. For example, a discussion in the working-age Sheffield group centred on increased hostility from neighbours following the Brexit vote, with one member commenting:

When migrant and minority ethnic individuals see the future of this country, they do not see themselves in it.

CP working-age group member

Outcomes of interest

Combining insights from the CP sessions with the theoretical literature, we identified the following types of outcomes, whether assessed quantitatively or described qualitatively, as being dimensions of loneliness, and therefore within remit:

• emotional loneliness, lack of intimacy

• social loneliness, (lack of) sense of belonging; feeling isolated

• feeling unsupported.

A framework of the proximate determinants of loneliness

Our theoretical review confirmed that existing formulations of the underlying causes of, and potential solutions to, ‘loneliness’ are inadequate. In particular, rather than pathologising people’s negative appraisal of existing social relations (employing terms like ‘abnormal’ cognition),2 it is important to acknowledge that individuals are exposed to social ties and interactions that undermine, rather than affirm, and that constrain, rather than facilitate, companionship. Importantly, for migrant and ethnic minority people, these interactions are commonplace, and threats to self-worth are routinely experienced. Furthermore, for many migrants, a loss or weakening of significant social exchanges is a prominent feature of their experience.

An initial scoping extraction of 49 interventions/initiatives involved mapping all of the inputs associated with these initiatives, and identifying the associated strategies (that is the cognitive, behavioural or environmental changes that were intended). We then examined each of these strategies, asking the question ‘how does this strategy have an impact on loneliness experienced at the individual level?’, to identify candidate proximate determinants that aligned with the theory review. We drew up a draft model and discussed it with the CP members. This process resulted in the identification of four proximate determinants, through which, we hypothesised, all initiatives must operate to affect loneliness:

1. increased positive ties and interactions – social relationships and exchanges that are experienced as providing affirmation and authentic companionship (being able to be oneself), often of a reciprocal nature

2. decreased negative ties and interactions – social relationships and exchanges that are experienced as failing to affirm, or as actively undermining, one’s sense of self, and/or failing to provide or actively obstructing valued social activities

3. increased self-worth – perception of being valued by, and valuable to, other people

4. more positive appraisal of existing social ties and interactions – reduced gap between what you want and what you have.

We recognised that these proximate determinants inter-relate with one another, and that there may be two-way causality (or feedback loops) between loneliness and these proximate determinants. For instance, experiencing negative social interactions is likely to undermine one’s self-worth. However, the model suggests that, in theory, interventions must affect loneliness via one, or more, of these determinants. So, for example, it may be possible to reduce loneliness by enhancing self-worth, even if there is no change to the number of positive social ties and interactions that the individual experiences. The model does not distinguish between intimate ties and broader social connections, whether experienced positively or negatively, because we found little indication that such a distinction would be helpful in understanding how interventions function. This decision also reflected a concern that this distinction may be ethnocentric, reflecting Western notions of the nuclear family and primacy of the intimate, heterosexual husband–wife bond. Work by Smart94 alerts us to the ways in which people are active in (re)creating family-type relationships, and that elements of intimacy, support and companionship can be achieved in different ways.

Potential strategies and solutions

Previous review work has categorised interventional strategies into four main types, each of which responds to an identified immediate cause of loneliness. 2 As shown in Table 4, we developed an extended model that enabled us to look for a wider range of potential interventional strategies.

Overview of the range of interventional approaches

We identified eight types of interventions. Table 5 lists these, along with their distinctive intended functions, and Table 6 shows the distribution of these types by the population subgroups we employed, as found in the published and more detailed grey literature (a total of 170 initiatives). Report Supplementary Material 1 provides brief descriptions of the intervention types found in the rest of the grey literature accessed.

In practice, not all of the interventions described in the published and grey literature conform exactly to the intervention types that we developed. In a few cases, interventions combine more than one of the intervention types (and several different functions). These ‘hybrids’ were assigned to the type felt to capture the most prominent distinctive function, with a note being added to describe their syncretic form. We also included two ‘other’ categories to accommodate interventions that did not fit neatly into any of the eight types. These were primarily multifaceted initiatives offering a menu of options to individuals or providing diverse activities and resources at a structural level.

In the next section, we provide a description of each of the identified intervention types, including their distinctive functions, along with an overview of the forms that they took, guided by the Template for Intervention Description and Replication (TIDieR)-Lite framework,95 when appropriate. Chapter 5 further develops the theory of how the common intervention types work and presents pictorial representations (logic models).

Befriending

The distinctive intended function 1 of interventions included in this type was identified as the creation of a one-to-one relationship of trust through which an individual with inadequate social connections receives tailored emotional support and companionship.

From our published (and more detailed grey) literature search, a total of 27 interventions were categorised into the befriending type, five of which were UK based (papers by McLeish and Redshaw96,97 that examined multiple interventions, some of which focused on our population groups, have been counted as a single intervention, as the analysis was aggregative). Diverse labels were employed to describe the individual acting in a befriending capacity, such as befriender, host, sponsor, tutor, buddy, labour friend, mentor mother and so on. Examples were found for all of our population subgroups except ‘children and young people outside educational settings’.

Appendix 7 sets out all of these interventions, describing the key components, guided by the TIDieR-Lite framework,95 and we summarise these here.

Shared-identity social support groups

The distinctive intended function 1 of interventions included in this type was identified as providing a safe, authentic, reciprocal social space where people who recognise some kind of shared identity engage with each other with meaning and enjoyment.

From our published (and more detailed grey) literature, a total of 49 interventions were categorised into this intervention type, eight of which were UK based (when papers drew on data relating to more than one support group but did not clearly distinguish between them in the findings, they have been counted as just one intervention each). Diverse labels were employed to describe the interventions, including support group, counselling group, self-help group, support network, (multi)cultural centre and others. Appendix 7 sets out the key components of each of these interventions as described in the papers.

Intercultural encounters

The distinctive intended function 1 of interventions included in this type was to bring together, in meaningful contact across ethnic/cultural/religious differences, people who do not normally interact.

From our published (and more detailed grey) literature search, a total of 35 interventions were categorised into this intervention type, 13 of which were UK based.

Training or equipping focused

Although other interventions, notably SSGs and befriending, could include an element of training or equipping, these interventions were distinctive in their focus on this function: to provide individuals with an opportunity to gain new knowledge and/or skills that have relevance to increasing the quantity and/or quality of their social ties and interactions. Eighteen interventions were identified as being of this type.

Psychotherapy

Eleven psychotherapy interventions were identified. The distinctive function of these initiatives in relation to our focus of interest was identified as follows: to provide individuals currently living with (or at high risk of) a common mental disorder with structured therapy to help them better understand and adapt their thinking and behaviours (which may contribute to isolation and loneliness). Eight out of 11 interventions were framed in terms of treating or preventing depression, whereas three had a broader focus on well-being in the context of challenging situations. 155–157

Meaningful activity focused

Although meaningful activity was a characteristic of other intervention types, notably SSGs, these interventions were distinctive in their focus on this function alone: to provide an opportunity to undertake activity that has meaning for the individual and the potential to enhance self-worth. Just five initiatives were identified as being of this type, although some of those typed as equipping or training could also be considered meaningful activity, such as tai chi classes. 152

Volunteering

One volunteering-based intervention was identified for older people, Experience Corps; this is a well-established initiative in the USA that involves placing older people in primary schools to support young children, and attracts a high proportion of African American participants. 174,175 Vickers176 reported on volunteering initiatives for people seeking asylum and refugees in UK. Two initiatives involved opportunities for minority students in the USA to engage in volunteering that showcased their language and culture. 58,177

Volunteering was also an element in some of the SSGs, and the intercultural encounters, particularly those reported in the grey literature.

Light-touch psychological inputs

Interventions of this type expose individuals to brief ‘messaging’ that is intended to alter perceptions and support more positive appraisal of existing social ties and interactions. Five initiatives of this type involved university students60,178–181 and two involved school children. 182,183 All were US based, and a majority of these involved the same group of researchers. In all cases, the focus was on increasing students’ sense of belonging – fitting in and being valued – in their educational institution, with the longer-term goal of supporting retention and academic achievement.

An example is an initiative that involved a 1-hour laboratory-based exercise intended to convey the message that social adversity in school is short-lived, and not attributable to fixed qualities of self or one’s ethnic group. Students were given the results of an ostensible survey of senior students about their experiences of college. They were then asked to write an essay and prepare to deliver a speech to new students that echoed the findings of the survey. The intention was that students would internalise the message and not experience subsequent ‘ambiguous’ social interactions as social setbacks or indicators of their lack of fit. 179

Other (individual)

We also identified a further five initiatives delivered at the individual level that did not fit easily within the eight types already mentioned. One of these was a social prescribing initiative involving referral to varied local programmes or activities and support from a volunteer tailored to individual needs. 184 Another was a US programme in which people living in deprived neighbourhoods were entered into a lottery that, for some, gave them the chance to move to a better-off area, with the potential for opportunities for better housing, education and employment, but uncertain impacts on social connections. 185,186 Strang187 evaluated the Holistic Integration Service for refugees in Scotland, which included a large number of different components, including some directly relevant to social connectedness. Franz188 described an outreach social work initiative that involved engaging first- and second-generation migrant youth in Austria in music and creative activities, with the potential to positively affect social connections and self-worth, and Kipling189 examined the role of citizenship ceremonies on integration and belonging.

Other (community, organisation or structural)

A further nine initiatives were identified that were judged to have relevance to social connectedness, particularly bridging ties and sense of belonging, for migrant and/or ethnic minority groups, but that operated at organisational and partnership levels, seeking to create an enabling environment, or to support delivery by other agencies. Papers that described these initiatives did not include exploration of individual-level processes or outcomes. These papers examined UK Children’s Fund strategies aimed at supporting social network-building among refugee and asylum-seeking children;190 refugee support organisations in Austria;191 a multicomponent organisation in Peterborough, UK, involving both direct service provision to refugees and a base for other community projects;192 a US action research initiative involving older people in shaping physical and social environments;193,194 an approach to social capital development among a US Latino community;195 an active learning pedagogical change initiative in a US university;196 a school-wide model that aimed to enhance prosocial behaviour;197 French local authority policies relating to accommodation of Roma migrants (villages d’insertion);198 and community-based organisations providing settlement services in South Australia. 199

Training or equipping focused

As noted in Chapter 4, several of these initiatives focused on equipping migrant or ethnic minority individuals with the sociocultural skills, as well as contextual information and navigational skills, perceived to be necessary to function well in a new/challenging role or social setting. 146,147 As a result, the logic of these initiatives mirrors the ‘equipping for context’ function found in befriending (function 2) and in SSGs (function 3). In a few cases, these interventions focused on equipping recipients with a narrower set of skills, such as learning to drive148 or improving balance,217 or took an approach focused more on health and well-being. 150,151 Nevertheless, in relation to our focus of interest, the logic of these initiatives can be depicted using the same causal chains. Those interventions that utilised a group-based format, either alone or in combination with individual elements, could also be represented as directly creating the opportunity for recipients to form new social ties with other group members (although this was not the primary function and was not always made explicit).

Psychotherapy

Interpersonal psychotherapy158,160,162 was described as aiming to improve an individual’s ability to cope with, adapt to and improve his/her social environment. This type of therapy has a problem-solving focus and addresses the ways in which individuals communicate their attachment needs to others, thereby aiming to facilitate the development of more supportive relationships. 160 ‘Cognitive–behavioural’ interventions155,159,161,164,165 similarly aim to enhance recipients’ awareness of social interactions, and improve social and communication skills. Jesse et al. 159 noted that:

[W]hile a CBI [cognitive–behavioural intervention] cannot compensate for lack of partner or other support in women’s lives, the women can use skill-building activities to help build a healthier support network, to communicate better with existing support persons, and to learn to ask for what they need/want.

Jesse et al. 159

An African-centred and spiritually centred psychotherapy group for black men similarly focused on understanding and improving social relationships within and beyond the family. 157 In an intervention for carers of people living with Alzheimer’s disease,156 structural ecosystem therapy was described as building up family relationships and assets, and restructuring specific interactions within the family, and between the family and other systems, that may contribute to caregiver burden.

The key causal chains hypothesised for most of these interventions operated via ‘improved interactions with existing ties’ to affect proximate determinant 2, that is to reduce negative ties and interactions. These initiatives were also expected to increase access to new social ties, and thereby operate on loneliness via proximate determinant 1. Intermediate changes that were hypothesised included reduced stress/increased trust and increased self-confidence/mastery.

Those interventions that utilised a group-based format, either alone155,164,165 or in combination with individual elements,156,157,159,161 could also be depicted as intending to directly create the opportunity for recipients to form supportive social ties with other group members (i.e. pathway 1a on the befriending and SSG models).

Meaningful activity-focused

In common with intervention types mentioned above, meaningful activity initiatives that were communal in nature (e.g. community gardens170) could be hypothesised as acting partly via the creation of positive social ties and interactions. However, this type of initiative was distinctive in that its core focus was on the individual engaging in an activity, through choice, that provides a sense of purpose and fulfilment, and often an opportunity to develop mastery and self-confidence. In addition, such activities may result in reduced rumination. These activities are hypothesised as acting on loneliness primarily via increased self-worth and via more positive appraisal of existing social ties and interactions.

Although not represented in our published papers, acts of religious observance, such as prayer, meditation and fasting, were identified by CP members as a further type of meaningful activity that can be supported by interventions, or indeed undermined, and are hypothesised as affecting loneliness via self-worth and an experienced positive connection to God.

Volunteering

The main hypothesised causal chains associated with volunteering relate to the provision of an opportunity for the individual to ‘give back’ and, in so doing, to demonstrate both competence and value, and to gain self-confidence. Increased self-worth is anticipated to result from this. Because volunteering will also often involve direct social interactions, as in the case of older people working in teams to support young school children,174 there is also the potential for a direct increase in positive social ties.

Light-touch psychological inputs

Interventions of this type are unusual in that they include no inputs that are intended to alter the number or quality of a recipient’s social ties and interactions (at least not in the short term). Instead, they are intended to operate via adjustment of the recipient’s cognition so that they develop a more positive appraisal of existing social ties and interactions (proximate determinant 4), and greater self-worth (proximate determinant 3), which then results in an increased sense of belonging. The primary assumed causal chain is via the normalisation of social challenges and/or reduction in internalised stigma.

Distinctive function 1: one-to-one tailored support

Our generic logic model identifies the distinctive function of befriending interventions as ‘provision of a one-to-one relationship of trust that provides tailored emotional support and companionship to the befriendee’. The assumption is that, if such a relationship is established, there is a direct improvement in the befriendee’s level of positive social ties and interactions (proximate determinant 1) via pathway 1a.

Additional function 2: equipping for context

Our generic logic model identifies ‘equipping the befriendee to function well in a new or challenging role or context (often linked to change or transition)’ as additional function 2 of befriending initiatives. This function was further described as providing the opportunity for beneficiaries to increase knowledge of relevant systems, norms, and processes; develop social and communication skills; and also develop practical, problem-solving and coping skills. All of the evaluative studies included here described befriender interventions that included some element of this function.

Additional function 3: linking to support

Our generic logic model identifies an additional function 3 of befriender interventions as ‘linking the befriendee to other social support opportunities and services’. This function was further described as potentially providing the opportunity for beneficiaries to receive tangible support (e.g. transport), advocacy, navigation and direct social introductions. Ten of the evaluative studies included here described befriender interventions that included some element of this function.

Feedback loops

The logic model identifies four potential positive feedback loops (see Figure 2). Although none of the studies reviewed focused on these relationships in any detail, some evidence in support of some loops was found. For example, McLeish and Redshaw’s96 findings described the way in which some peer supporters introduced groups of mothers to each other and suggested that this could contribute to a virtuous circle of increasing confidence and social connectedness (feedback loops 1 and 4). Similarly, Askins202 described how, over time, some befriendees became more confident and able to engage in new connections, which, in turn, led to greater confidence and firmer relationships, so that they became ‘progressively more involved.’ 202

Second-order causal links and proximate determinants

The logic model suggests that, as self-confidence/empowerment/mastery increases, this should lead to befriendees accessing wider opportunities for social support and interaction (pathway 4), as well as improved interactions with existing ties (pathway 5). Similarly, the logic model suggests that increased trust and reduced worry/stress should also lead to befriendees accessing wider opportunities for social support and interaction (pathway 6), as well as improved interactions with existing ties (pathway 7). We found some qualitative evidence in support of these pathways, but it was often difficult to disentangle pathways, as increased confidence/mastery was often presented as interplaying with increased trust and reduced worry in these qualitative studies. For instance, McLeish and Reshaw96 concluded that befriending schemes often had ‘substantial and interlocking positive impacts on emotional wellbeing’. 96 Improved interactions with family ties was identified as an important intermediate effect of several initiatives:99,200,201,218

By talking things over with my family support worker it has been a lot of help to us all. Before, he was beating me and my kids had physical abuse . . . But the worker talked to him . . . and he thinks it has helped. He used to call her his sister.

Gray99

Evidence from Askins202 suggests that, for some individuals who are seeking asylum, developing trust is a long-term process and persistent mistrust can impede the development of new social ties. These findings suggest that equipping such newcomers with the knowledge and skills to operate in their new home is not sufficient to lead to the development of new close social ties, nor necessarily improved interactions with existing ties, thereby demonstrating that these pathways are distinct.

Outcomes

Nine qualitative evaluations demonstrated reduced feelings of isolation or loneliness, and/or increased feelings of belonging, or of being supported, among a majority of befriendees in receipt of the interventions. However, all involved data collection during, or soon after, the end of the intervention period, so that they provide no evidence relating to longer-term impacts on the review outcomes of interest.

Three quantitative studies reported some kind of positive effect on a relevant outcome. Two evaluations of adult mentor programmes in schools reported an increased sense of belonging to school among children. 222,226 There was an increase in feeling supported among women in receipt of a Promotoras intervention. 229

Four quantitative studies found no effect on review outcomes of interest. Two of these224,228 reported no difference among the recipients in measures of feeling supported. One228 of these also reported no effect on loneliness [using the University of California, Los Angeles (UCLA) Loneliness Scale57]. Low et al. 223 reported no effect on loneliness (using the UCLA Loneliness Scale57) of an intervention among elderly recipients and Quintrell and Westwood221 found no difference in the proportion of international students who described their first year at university as ‘lonely’ between those who had and those who had not been part of a host support programme. Importantly, none of these studies demonstrated pathway 1a, suggesting that they failed to achieve the distinctive function of befriending, namely one-to-one tailored support. We return to these quantitative studies later when we consider some generic hypotheses about likely effectiveness.

Differential processes and outcomes

Few of the studies reported on differential outcomes or considered intersectionality. There was evidence of a gender difference in the intervention’s impact on self-worth in one of the school-based initiatives, with only boys benefiting. 226 However, a majority of the befriender initiatives reviewed were targeted at women only.

Achieving functions: what inputs are important?

Befriending: negative effects

Two types of potential negative effects were identified by a small number of studies. The first related to the potential for befriendees to become dependent on the befriender, and therefore to be at risk of feelings of loss when the relationship comes to an end. 97,201,224 These studies highlight the need for careful management of emotional relationships, boundaries and expectations, as well as careful consideration of how endings should be orchestrated for particular people in particular circumstances. The second negative effect was reduced self-worth (labelled as ‘self-esteem’) reported in just one study. 228 Previous work has emphasised the importance of selecting and matching befrienders (or mentors) carefully so as to avoid a relationship that undermines, rather than boosts, self-worth. This study involved no such matching and provided online, asynchronous interactions only.

Distinctive function 1: safe reciprocal space

Our generic logic model identifies the distinctive function of SSG interventions as ‘to provide a safe, authentic, reciprocal social space in which people who recognise some kind of shared identity engage with each other with meaning and enjoyment’ (shorthand: ‘safe reciprocal space’). The assumption is that if such a shared space can be created, there is a direct improvement in the participants’ levels of positive social ties and interactions (proximate determinant 1) via pathway 1a.

Additional function 2: buffering hostility

Additional function 2 of social support groups was identified as ‘buffering the external hostile context that denigrates ethnic minority and/or migrant identity’. Fifteen of the evaluative studies63,101,102,104,109,177,208–210,213–216,230,233,236–240 included described support group initiatives that included some element of this function.

Additional function 3: equipping for context

Additional function 3 of support group interventions was as follows: equipping group members to function well in a new or challenging role or context (often linked to change or transition). All but three of the evaluative studies included here described support groups that included some element of this function.

Feedback loops

Few studies explicitly explored potential feedback loops, but some evidence was available indirectly via participant stories. Some examples are given here.

Feedback loop 4 (see Figure 3): refugee women exposed to a peer support group plus mobile phone initiative were reported as becoming ‘more confident about travelling to places outside their home or community, which led to feelings of empowerment and self-capability among the women’. 234

Some papers also suggested positive feedback loops between the sharing of knowledge and advice within the group, growing confidence, and the experience of positive social ties. That is, the more group members share, the more they grow in confidence and the more they feel connected. 210,235 This concurs with the theory presented in Chapter 3 that the performance of social interactions/exchanges and the meaning of relationships are co-constituting.

Second-order causal pathways and proximate determinants

Outcomes

Evaluations of 19 interventions provided qualitative evidence that members had reduced feelings of social isolation or loneliness or developed a sense of belonging as a result of participating in the group:

What I gained was, you know, the sense of belonging.

Pejic235

Notwithstanding poor study designs, six studies103,214,230,237,238,248 included some quantitative evidence in support of an effect on relevant outcomes, whereas three63,101,109 provided quantitative evidence of no such effect.

As with the befriending initiatives, no studies involved long periods of follow-up, so none was able to offer evidence on whether or not positive outcomes were sustained.

Differential processes and outcomes

Most studies did not report on differential outcomes between subgroups nor engage in any analysis of intersectional processes. In two studies, focused on newly arrived refugees, authors found that the development of social networks by participants was dependent on pre-existing networks and assets, with some groups starting from a much lower level and developing more slowly. 214,215,243

Canuso68 reported that the group format did not work for two participants who had significant additional difficulties.

Saito et al. 103 reported that a programme for older people was effective on the loneliness outcome (as well as subjective well-being) for those who had at least mild loneliness at the start. In contrast, those who had no loneliness at baseline showed increased familiarity with services but no change in subjective well-being or loneliness. 103 Another programme for older people showed the greatest reduction in loneliness among ethnic minorities. 248

Two studies examined subgroup differences but found no effects, between Hispanic and black students233 and between immigration status groups. 232

Achieving functions: what inputs are important?

In this section we summarise the evidence that was available in relation to how particular inputs (the nature and volume of resources and activities provided) influenced whether or not the groups achieved the intended functions. Chapter 7 presents information on factors that influenced the access and uptake of initiatives.

Shared-identity social support groups: negative effects

Three potentially harmful effects of support group participation were identified in the papers: disclosure of sensitive information by group members within (and possibly beyond) the group; disruption to family relationships; and the potential to inhibit ‘out-group’ interactions, leading to further exclusion from wider society. There was very limited evidence in support of any of these effects; however, most papers did not include any discussion of unanticipated or harmful consequences of the initiatives.

Distinctive function 1: meaningful interethnic contact

Our generic logic model identifies the distinctive function of intercultural encounters as being to bring together, in meaningful contact across ethnic/cultural/religious differences, people who do not normally interact. This is further described in terms of providing an opportunity to share stories and experiences (listen and be heard), demonstrate competence and value, engage in purposeful activity and engage in enjoyable social activity.

The level of detail provided in the papers was often limited, and some papers examined multiple different initiatives under the umbrella of a programme, so that it was not possible to examine the elements of function 1 in detail across most papers. Several interventions did include explicit aspects that enabled minority participants to demonstrate competence,110,111,115 and, in some cases, leadership. However, a feeling of unequal value among minority participants was shown to persist in some cases. 112

Additional function 2: myth-busting

As noted above, function 2 and function 1 could not easily be disentangled in several of the initiatives, and the causal pathways were not generally clearly articulated in the papers.

Four studies112,113,126,130 provided qualitative data in support of pathway 2a, that ‘myth-busting’ could lead to more positive ‘other-group’ orientation among minority participants, and two studies112,113 provided evidence that ‘myth-busting’ could lead to more positive ‘other-group’ orientation among majority participants, at least in the short term.

One study137 indicated that myth-busting activities were not effective in shifting ‘other-group’ attitudes among majority participants, although these were not delivered in the context of an initiative that had clearly achieved function 1. The studies reported by Goodkind et al. 113 and Goodkind254 that evaluated ‘learning circles’ for students and refugees with two different refugee groups, Hmong and African, also both provided qualitative support for pathway 2b, that myth-busting can lead to increased trust and reduced worry.

Additional function 3: equipping for contact

Just seven interventions included some elements that could be considered to be aimed at function 3, and all of these were focused on people seeking asylum or refugees or other newcomers. In three studies,112,115,122 qualitative evidence was provided to suggest that pathway 3a was achieved, with migrant participants showing increased self-confidence/mastery. Participants in a football-based initiative reported that the leadership training provided had importantly increased their self-confidence:

The most powerful experience I’ve ever had in my life is that leadership program, you know, that changed my life, that changed my thinking . . . you know, make good decisions . . . they give you a good opportunity for you to be a good leader in the community (FS, male).

I: You’re officially coaching . . . how does that feel for you? R: It feels amazing. I look after people, then the people listen to me in the game and stuff (FS, male).

Nathan et al. 115

This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (https://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited

None of the papers provided convincing evidence for or against the causal chain between ‘equipping for contact’ and increased trust and/or reduced stress or worry, reflecting a lack of exploration of these potential processes.

Proximate determinants

Outcomes

Nine studies72,111,113,123–125,127,128,253 provided some data relevant to review outcomes of interest. Four studies72,111,124,253 provided some qualitative evidence, and three113,123,128 provided quantitative evidence, of an increased sense of belonging, feeling supported or reduced feelings of isolation. One of these qualitative studies124 reported mixed findings, with the initiative not being experienced as supportive by all, and two111,253 provided very limited data. One study127 drew on qualitative and descriptive quantitative data and concluded no evidence of improvement in a relevant indicator of belonging. A further study showed no evidence of improvement in a quantitative measure. 125

Differential processes and outcomes

As noted above, several studies noted that interventions did not play out similarly for all participants, or across all sites, but only a few identified systematic differentials between subgroups or contextual characteristics. Borgogni and Digennaro111 noted the low participation of women, and thereby an inability to evaluate the impact on women of an Italian football-based initiative. Mayblin et al. 142 concluded that higher education, and having foreign language skills, were factors that made it easier for some participants to engage in ‘organised encounters’ offered via a Polish football-based initiative. Nathan et al. 115 found that the effects of a football-based initiative for young migrants were similar regardless of how long the young people had been in Australia. Raw’s127 evaluation of a UK school-linking project identified multiple differentials: greater impacts among BME children in terms of other-group orientation and interethnic friendships, particularly BME girls, than white children; strongest effects among year 4 children; and stronger effects among children attending ethnically homogeneous schools than mixed schools. Meanwhile, socioeconomic status and faith versus secular school did not appear to affect outcomes.

Achieving functions: what inputs are important?

As noted above, several papers provided only brief information on the nature of the interventions, so that our ability to isolate key characteristics of inputs that are necessary or sufficient to achieve functions is limited. Nevertheless, the findings that are available tend to support earlier work91 that has characterised the nature of ‘meaningful encounters’.

Intercultural encounters: negative effects

A few studies suggested that, in some circumstances, intercultural encounters could exacerbate negative attitudes, inequalities in status and racist behaviour. Raw127 found that, in some schools, there was an apparent hardening of negative attitudes towards out-group members, particularly among majority ethnic children. Askins and Pain110 found that when children were involved in a photography project that gave them little active involvement, and instead positioned them as objects, ‘other young people’s comments were predominantly negative, based on (perceptions of) physical difference – the quintessential marker in constructions of difference – leading to arguments or evasion.’110

Goodkind256 found that some Hmong participants did not recognise themselves as imparting knowledge to the white American students they engaged with, instead positioning themselves has having less value (perhaps suggesting that the initiative heightened this negative self-perception, at least for some):

But I don’t think I taught you anything else because I learned more from you, because you know English, and you know how to read and write, and so I think I learned more from you than you learned from me.

Goodkind256

Pathways to increased positive social ties and interactions

There was some evidence that training/equipping initiatives could increase positive social ties and interactions for participants, either directly if they adopted a strong group-building element (and so, it could be argued, were similar to SSGs),75,257,258 or indirectly if they served to overcome important current obstacles to social interactions. 148 For instance, one paper illustrated that, for older individuals, a training-based intervention designed to reduce the fear of falling significantly reduced reports that a concern about falling had ‘interfered with normal social activities with family, friends, neighbours, or groups’. 217

However, interventions that focused narrowly on equipping participants in some way did not necessarily increase positive social ties. An intervention that aimed at equipping students with intercultural skills reported a small increase in a measure of cross-ethnic social self-efficacy, and an increase in time spent with cross-ethnic friends, but no increase in the number of cross-ethnic friendships. 147 An equipping intervention for older migrants in Sweden, despite aiming to encourage group members to discuss and share experiences, showed no positive impact on positive social ties. 153

Meaningful activities that were communal also showed evidence of increasing positive social ties,170,172,173 as did some volunteering activities. 259

Pathways to reduced negative social ties and interactions

There was evidence from two studies157,164 that interventions that employed psychotherapy could enable participants to develop the skills and confidence to better manage stressful relationships, leading to fewer negative interactions:

Yes, it has. I am managing my family with less stress; I have realized and have gained the knowledge of overcoming my tension when I am dealing with my children and husband. I tell them I can do only so much at a time and they shouldn’t be expecting a lot from me.

Masood et al. 164

This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (https://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (https://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated

Pathways to increased self-worth

There was evidence that meaningful activity initiatives could provide participants with a sense of being competent, of value and of being valued by others. 170,172

Volunteering activities could also increase feelings of self-worth:58,177

I believe it made me feel needed and important. It made me realize that I could use my abilities to help others and made me feel good about myself. The feeling of satisfaction was the most beneficial aspect for me.

Pak58

I feel like this [the SL (service-learning) project] gave Latinos a way to be involved and know that they can actually be a part of something.

Pak58

There was evidence from some interventions that employed psychotherapy that these could lead to enhanced self-worth (labelled as ‘self-esteem’),159,164 and that the active listening and sharing components of these group-based interventions were an important element (in common with befriending and SSG initiatives). 164

There was also evidence that some training/equipping initiatives could enhance self-worth when they successfully equipped participants with the ‘know-how’, confidence and skills that they perceived they needed to operate in an important context. Such initiatives often incorporated a peer-learning approach and involved a normalisation pathway. 75,151,260 Initiatives that failed to equip participants with the know-how and skills they wanted were less successful. 261

Pathways to more positive appraisal of existing ties and interactions

There was evidence from one study that an intervention that employed psychotherapy could enable participants to re-appraise their existing ties in a more positive light. Elligan and Utsey157 reported that participants demonstrated ‘an increased appreciation and love for other African American men’ and greater sense of tolerance for conflict in interpersonal relationships and a renewed interest to maintain intimate/romantic relationships.

Hypothesis 1

Loneliness (feeling isolated, feeling unsupported) has a complex aetiology and the underlying causes can vary between individuals. Therefore, interventions that target underlying causes of loneliness are more effective, on average, than those that do not. Three groups of interventions are likely to be effective; these are interventions that:

1. focus on a particular group of individuals in a particular setting or circumstance and are designed to address their particular needs

2. are flexible enough to enable attention to individual needs

3. involve multiple components that link to all, or most, of the proximate determinants.

Befriending initiatives might be considered the most flexible, because they are provided on a one-to-one basis. However, the befriending initiatives examined in the quantitative batch of studies tended not to demonstrate achievement of one-to-one tailored support. Four of these initiatives were unsuccessful. 221,223,224,228 Three papers concluded that the intervention was successful in relation to a relevant outcome: two222,226 focused on schoolchildren, a targeted group that was provided with inputs tailored to a child’s social and academic achievement in school, and the third initiative focused on Spanish-speaking women who were provided a linguistically and culturally tailored input from Promotoras. 229

Four studies examined interventions that provided multiple components aimed at addressing at least three of the proximate determinants. Three of these were reported to be successful. 113,214,230 The fourth63 was not successful in affecting quantitative outcomes, although there was promising qualitative data. The two interventions evaluated by Stewart et al. 63,214 were very similar, suggesting that contextual differences may have moderated the causal processes.

Hypothesis 2

Low self-worth (linked to wider hostile environments) is common among migrant and ethnic minority people. Therefore, interventions that explicitly aim to boost self-worth are more effective, on average, than those that do not. Two groups of interventions are likely to be effective:

1. interventions that integrate components that directly enhance self-worth (e.g. opportunities to reciprocate, to demonstrate competence, to build confidence via new skills)

2. interventions that enable user involvement and/or ownership.

Overall, the data provide fairly consistent support for this hypothesis. Six studies113,128,214,222,226,237 reported a positive impact on self-worth (quantitatively or qualitatively) and a positive impact on a relevant outcome measure; two further studies182,229 had an explicit intention to affect self-worth (although there were no data presented) and the findings suggested a positive effect on a relevant outcome. In addition, four initiatives did not explicitly mention self-worth, but either supported active user involvement248,262 and/or had a focus on a closely related concept: mastery211,248 or self-efficacy. 103,152

Three interventions had no explicit intention to improve self-worth and did not achieve significant impacts,109,125,146 and one had an intention to improve self-worth but did not appear to include relevant inputs and had no impact on outcomes. 263 A further initiative actually reduced self-worth (despite intentions to boost it) and had no positive effect on feeling supported or on loneliness. 228

In two studies,63,101 there was evidence of increased self-worth, but no associated improvement in outcomes; both focused on people seeking asylum and refugees. In Adam et al. ,101 there was no active involvement of users and there was evidence of unaddressed negative social interactions. In Stewart et al. ,215 there was qualitative evidence to suggest some improvements in perceived isolation and perceived social support.

Two interventions appeared to achieve a positive effect on outcomes despite no explicit attention to increasing self-worth. Both of these focused on international students and employed less rigorous study designs. 144,264

Hypothesis 3

For new migrants, particularly forced migrants, a severe lack of social connections is an important risk factor for loneliness. Interventions that explicitly support the building of a shared-identity social network are more effective than those that do not.

Three studies, focused on migrants, that put an emphasis on building shared identity ties reported an increase in positive social ties and a positive effect on a relevant outcome. 113,214,237

Two befriending interventions did not appear to aim for shared identity ties, and produced no impact on participants’ reports of positive social connections, nor on the final outcome. 221,224 Two initiatives, focused on international student adjustment, included no apparent emphasis on shared identity ties and produced no impact on final outcomes. 125,146

Two studies63,101 did appear to focus on shared identity ties, and did provide qualitative evidence of an increase in positive ties, but reported no impact on loneliness, suggesting that other factors were at play. Notably, in the Adam et al. 101 initiative, the target group was migrant LGBT+ men, and there was some qualitative evidence that experience of negative social interactions was not reduced (see Hypothesis 4).

Two studies144,264 provided contradictory evidence, in that the interventions did not appear to aim to increase shared identity ties but, nevertheless, reported improvements in relevant outcome measures for international students. It is worth noting that, in the Brunsting et al. 264 intervention, the participants were 86% Chinese students, and so may have developed shared identity ties even though this was not an explicit aim. This initiative was intended to operate via a shift in international students’ appraisal of the adequacy of their social ties, although no data relating to this causal pathway were provided.

Hypothesis 4

Having social ties and interactions that are experienced negatively is common among migrant and ethnic minority people. These can be (1) stressful intimate ties (due to living in conditions of material deprivation etc.) and/or (2) exposure to hostility and microaggressions in varied social settings, including with service providers and ‘agents of the state’. These negative social interactions can undermine the effect of positive social ties, and create loneliness. Interventions that fail to address persistent negative social ties and interactions are less successful than those that do address them.

Few of the quantitative studies explicitly engaged with these causal processes.

Two initiatives, focused on people seeking asylum and refugees, included an explicit focus on reducing negative interactions and provided some evidence of success in this regard, as well as reporting a positive effect on outcomes. 214,237 A further initiative, focused on schoolchildren, also provided some evidence of a reduction in negative interactions and reported a positive outcome effect. 222 Four further initiatives had an explicit aim to reduce negative ties and interactions, and reported a positive impact on outcomes, although no data were provided on this proximate determinant. 113,155,158,160 One initiative aimed to affect this determinant, but appeared not to be successful and also reported no impact on the loneliness outcome. 101 One study failed to affect the outcome despite some evidence of positive ties within the group; qualitative data suggested that this was because negative intrafamilial ties had not been addressed for many of the women. 165

One initiative recognised the importance of addressing negative ties, but provided no data on this determinant (and did not appear to provide credible inputs to address it), and reported no effect on the outcome. 228 A further initiative, for international students, recognised these factors and reported no effect on a relevant measure, but did report an impact on the outcome. 264 An initiative for refugees reported some qualitative data of impact on negative ties, but no effect on final outcome. 63

Twenty-two initiatives provided no information to suggest that they recognised or targeted negative ties and interactions. Ten of these reported no impact on final outcomes. 60,109,125,146,159,221,223,224,263,265 Twelve initiatives reported a positive impact on a relevant outcome,103,128,144,150,152,179,182,226,229,230,248,262 although, as already noted, study quality was often poor.

Hypothesis 5

For individuals who have multiple, interlocking risk factors, addressing loneliness will take time, and short-term improvements may not be sustained. Interventions that are longer term, or that include effective transition arrangements to connect individuals with secure follow-on sources of support, are more likely to be effective than those that are of a fixed, shorter duration.

Quantitative studies did not include sufficiently long durations of follow-up to test this hypothesis.

Self-confidence; self-worth

A theme of low self-confidence undermining people’s ability to form and maintain social connections was common across all CP groups. CP members across all groups also identified important feedback loops. Supportive relationships serve to enhance self-worth, which, in turn, encourages more interaction and the chance to develop new relationships:

Boosting the confidence is the key, it also gives them a feeling of sense of worth – they are not useless, they are not old . . .

CP older persons group

Being confident in your heritage and can go out in the world and say that ‘this is me, my heritage’ and you can go and show this.

CP student group

The student CP members felt that self-worth was a ‘defining thing’, and noted that a lack of self-confidence could prevent people from taking up the opportunities around them to engage socially. They also noted how contextually sensitive self-confidence and self-worth could be, and how the lack of supportive people around a person could quickly undermine these feelings.

Low levels of confidence could prevent people from taking up available offers of support. A CP member shared a story of a young male asylum seeker who was fearful of attending a youth group and needed the safety of one-to-one supportive chats over a period of time before being confident, and trusting, enough to engage with the group. Askins202 found that some people seeking asylum were not even confident enough for a one-to-one befriending service.

Several closely related factors were identified, via the CPs and published literature, that feed into low confidence and undermine self-worth.

Ill-health and disability

The relatively high levels of chronic ill-health and disability among ethnic minority groups, particularly older people, was identified as an important factor that both contributes to isolation and loneliness and undermines people’s ability to take up interventions. Again, this theme appeared in both published studies and CP discussions.

The negative spiral of depression, leading to social isolation, leading to more chronic depression was highlighted in CP workshops. CP members described how people can ‘become stuck’, and breaking out of depression, isolation and low self-worth can be very difficult.

Physical ill-health among older people was felt to be a particular issue, and the importance of outreach services was highlighted. The Irish centre evaluated by Cant and Taket104 catered to people who felt particularly isolated and unable to visit the project because of health problems by providing a culturally sensitive telephone befriending service.

Material resources

The impact of limited material resources on people’s ability to engage in social activities, both informal and of a more structured provision, was mentioned across all CP groups. Students talked about being invited to go to the pub, but not having the money to do so. Other CP members talked about people seeking asylum not being able to participate in activities that were intended to connect them to others, such as free swimming sessions when not having the money to buy swimming trunks. Lack of money was a particularly prominent theme for forced migrants, in both the CPs and the published literature, with some not even being able to accept an offer of going for a cup of coffee with a friendly neighbour. Furthermore, the expectation of having to pay can deter people from venturing into community centres and other spaces,69 and having to ask for favours (e.g. eating with friends when awaiting benefits) can put a strain on social relationships. 187 There were numerous examples of how interventions can fail to adequately consider the resource constraints that people face. 131,202 In Askin’s202 study, befrienders had to work around the lack of resources experienced by asylum-seeking individuals by meeting in befriendees’ houses, as this involved no direct cost such as bus fares or buying drinks.

Awareness and familiarity

Consultation panel members noted that older people and newer migrants may have limited awareness of opportunities for social connection and support, and may be reluctant to venture too far afield. There was a feeling that many people rely on friends and family for information and recommendations about places to go and things to do. Places can feel unfamiliar and threatening both to new arrivals and to established residents as their character changes over time.

Low uptake of interventions was reported by CP members and in published papers. Word of mouth, and working through people and organisations already known to intended beneficiaries, can be helpful in increasing participation over time, and breaking down mistrust and/or scepticism. 110,266 Respectful partnering with local community organisations to reach ethnic minority people had been successful in some locations. 267 In the case of some initiatives, there may also be a need to explain unfamiliar content to prospective participants. For example, Pakistani women with depression who were invited to join a psychotherapy-based support group described their initial apprehension about the nature of the groups and the degree of disclosure that would be expected. 268

Individual needs and preferences for social relationships and contact

Consultation panel workshop discussions often focused on the ways in which particular interventional forms would not be suitable for everyone, highlighting the range of individual attributes and circumstances that shape needs and preferences for social contact and relationships.

Faith and spirituality

Faith (as opposed to religious identity) as a resource that can protect against loneliness was identified by members in two of the CPs. There were no UK intervention papers that explicitly paid attention to this aspect, although some US-based initiatives did. We are aware of initiatives beyond the scope of this project that have been developed in the UK to allow the integration of people’s faith (and personal religious resources) into recovery from depression (although these are not directly concerned with loneliness). 271

Loss of family ties and support

All of the CPs talked about family ties as being foundational to a person’s sense of being supported and belonging. Student CP members talked about the difficulties they faced being a long way from family members who were an important source of emotional and affirmational support. Efforts were made to keep in touch through social media, but time differences and virtual contact did not feel the same as face-to-face interactions for some. These CP members also felt that some relationships were lost, or weakened, irreparably with the passage of time. Loss of family was also a salient theme in the published studies relating to forced migrants:

No, I do not think that here is like Albania, because there are so many things . . . . I miss family.

I never lived alone, I am coming from a family where I have five sisters and I always surrounded people, and I was scared of living alone.

Refugees, Strang et al. 187

Reproduced with permission from the Scottish Refugee Council

Older CP members also talked vividly about the permanent loss that results from being a long-term migrant. This sense of loss related not simply to loss of social relationships, but to a loss of place:

It’s not there anymore – it has changed, no longer the close family. Once the parents have gone, you go back to visit and they’re not there. House has gone, and no one to share and no memories are left.

CP older persons group

Despite loss of family ties, there was also evidence of active agency on the part of forced migrants. Clayton272 looked at young ethnic minority people in Leicester and found that they were active in maintaining both local and transnational ties that were important in their sense of belonging; multifaceted identities were not necessarily in conflict, and their religious identities were important. 272

Recognition of the importance of maintaining transnational ties, and support to do so, appeared to be limited in UK interventions. A pilot initiative in Sheffield identified by a CP member sought to equip older Chinese people to use smartphones and WeChat (Tencent Holdings Ltd, Shenzhen, China) so that they could connect with relatives in and outside the UK.

Families as sources of stress; negative social support

Both CP discussions and published literature highlighted the importance of recognising that family relationships were not necessarily supportive for migrant and ethnic minority people. Furthermore, given expectations of close family ties and reciprocal obligations, poor family relationships could often be felt very acutely:

But, there is an expectation of having to please your parents. There can be a feeling that you are not pleasing your parents and not living up to their expectations. This can undermine sense of support. In this situation, individuals may be going out and about and connecting with others, but they may still feel isolation and loneliness if parents don’t give support to them.

CP working-age group, talking about South Asian women experiencing divorce

In the papers describing psychotherapy and group-based support for women, negative support from spouses, conflict with children and unfulfilled expectations of support were common themes contributing to feelings of isolation and loneliness. 201,218,250,252

Consultation panel members who were migrants, both students and those who were working, talked about the way that interactions with family members and close friends ‘back home’ were not necessarily supportive:

If my mum says to me when I’m feeling frustrated and sad ‘Come back to [name of country]’, it is very different to her saying ‘Oh don’t worry, you’ve just had a bad day, why not go out for a walk?’.

CP working-age group member

People could be judgemental, have high expectations and make assumptions about the migrant ‘having an amazing time’. This could make migrant individuals less inclined to keep in contact with their relatives and encourage them to conceal their need for support. CP members noted that such expectations can make people feel that they are not achieving, undermine self-worth and increase feelings of isolation:

You are torn between the two. . . . I put it as a negative sort of connection. If you feel that connection is there, you feel you have got that responsibility and the perception is you’re scared, you aren’t there – land of opportunity, but you are torn – you long for it, you want to hear your mother’s voice, but you try, you know, you are torn.

CP student member

There were also disappointments at not being able to share new experiences with close friends:

My girlfriend back in Mexico – I didn’t feel lonely by this time – but I was sharing all my new experiences here and for her it was not that exciting because she was in the same routine. I felt lonely because she wasn’t sharing with me, this feeling or this emotion.

CP student member

Family members were also found to restrict access to interventions in some cases. Masood et al. 164 reported mixed support for South Asian women to attend a group-based initiative, and Gater et al. 268 identified resistance from family members, particularly husbands, as a ‘major hindrance’. 268 Gater et al. 268 noted that family members ‘did not recognise depression as meriting outside help and lacked faith in the appropriateness of the intervention’, indicating the importance of gaining trust and providing culturally sensitive care.

Family members could also discourage positive cognitive and behavioural shifts among people who were participating in projects aimed at creating positive connections. The Respect and Understanding – Building Inclusive Communities (RUBIC) project116 was described as a multilayered model that tackled attitudes in students’ homes, as well as involving students in leadership activities, as this was anticipated to have ‘a much deeper impact’ than working with young people alone.

Failure to address sources of negative social support may mean that interventions are not successful, as the provision of positive ties may be insufficient to counteract negative ones if they are significant relationships about which the individual holds particular hopes and expectations. However, relatively few of the UK initiatives appeared to actively engage with these issues.

Competing demands and responsibilities

Competing responsibilities and demands were identified as undermining people’s abilities to invest in their social relationships, as well as their access to and engagement with interventions.

Housework and child-care responsibilities were highlighted for women. For instance, ‘homework’ expectations of a group-based psychotherapy initiative were difficult for South Asian women with child-care responsibilities to complete. 164 Similarly, refugee women with children found it difficult to combine child-care responsibilities with the timing of English classes. 187

Consultation panel members suggested that men from migrant communities were also at risk of being socially isolated owing to their felt duty to earn for their families. For instance, many male members of the Chinese community were identified as working long unsociable hours in the restaurant trade, leading to being isolated not only from family, but from the wider community.

Interventional approaches could increase participation by providing child care, and convenient scheduling. In addition, both CP members and published evidence suggested that the content of initiatives is important, in that people may feel able to prioritise, and justify to other family members, activities that provide clear practical skills and knowledge that have value beyond the individual.

Stigma

Felt and enacted stigma relating to feelings of social isolation and loneliness were highlighted in all CP workshops. As noted above, migrant individuals could feel reluctant to admit to family members ‘back home’ that they needed support, when they were perceived to be the lucky ones.

Consultation panel members felt that men felt particular stigma to admitting feeling unsupported or lacking in social connections, with gendered expectations forcing them to ‘man up’ and ‘suffer in silence’.

At a family level, there can be resistance to individuals, particularly older people, accepting support or engaging in initiatives, because it reflects badly on other family members who have a perceived obligation to meet those needs themselves.

Consultation panel members suggested that initiatives that enable participants to involve their wider family members from time to time, for instance a gardening project, may be more acceptable and also provide a way for the individual to demonstrate competence and to strengthen important family ties.

Interventions that adopted neutral labels, for example ‘Thursday night chat’, or that were framed in terms of aspects of development could be less stigmatising. As an example, international students may feel embarrassed and illegitimate in devoting time to a support group that takes time away from their studies, but may feel able to join a group that is focused on adapting to life in a new academic environment (which at the same time provides emotional support and companionship).

Feeling unwelcome; interpersonal racial harassment

There was a strong and consistent theme of feeling unwelcome:

I feel lonely everywhere, not just in London. People don’t like migrants, especially if you have the face of an Arab or a Muslim . . .

I felt remarkably unwelcome in the UK. There was a distinct feeling that you, as a ‘foreigner’ were not welcome . . .

Refugee participants, Christodoulou. 239

Reproduced with permission from Migrants Organise

More seriously, direct experience, and fear, of racial abuse as something that limits people’s social interactions and contributes to feelings of isolation and lack of belonging among both migrants and ethnic minority people was a consistent theme across all CPs and the published literature. In a vivid example, a member of our older people’s CP described how someone spat in her face in a local area as she was dressed in her traditional clothes. In another, they described newly arrived migrants in one area of the city running to the bus stop as quickly as possible to evade the torrent of verbal abuse from neighbours. In Gray’s99 study, the young and the elderly in the Bangladeshi community were found to feel especially vulnerable, with respondents citing multiple incidents of theft and bullying:

I had a mugger who took my money. [. . .] The mugger hit me as well, so I reported it to the police like my family support worker told me to. [. . .] Now I’m afraid to go outside myself and I’m afraid for my son to go outside alone.

Gray99

The risk of abuse was exacerbated by poverty. For instance, children attending school in old clothes became a target owing to their ‘visible poverty and cultural differences’. 99

A CP member recounted experiences of neighbour abuse:

I know that when we moved into a bigger house in [name of place] in the centre, our neighbours weren’t happy from next door, that a Pakistani family has moved in; ‘Our house prices will go down’. They did everything possible to offend us, sort of thing.

CP working-age group

Experiences of racial abuse and harassment were identified as a particular problem for those seeking asylum and refugees. In part, this was identified as being related to having no control over their housing and being housed in areas where other residents hold racist attitudes. In addition, however, the overwhelmingly negative portrayal in the media and politics was felt to be a significant contributing force.

Furthermore, although CP members recognised that skin colour, religious dress and other aspects of appearance can make some migrant and ethnic minority people particularly visible, fear of racism was common among white European migrants also:

I know of someone from Polish origins who told her children not to speak in Polish outside because then they will become easily identifiable.

CP working-age group member

Consultation panel members felt that the prevailing negative racial climate limited people’s knowledge, and uptake, of opportunities to engage in activities and programmes aimed at tackling isolation and loneliness. They also expressed concerns that progress towards building self-confidence and a sense of belonging that might be achieved in group-based activities, and other initiatives, can be undermined by negative experiences out and about – themes that resonate with the published evidence on intercultural encounters that are reported in Chapter 6.

Although some UK initiatives that were reviewed explicitly engaged with racism, recognising this as part and parcel of migrant and minority experiences, this was uncommon aside from those labelled as intercultural encounters. Notably, recognition of the need to address racism was more prominent in SSGs in the USA than in those in the UK.

Neighbourliness

Importantly, as well as revealing a negative focus on racism, CP members and the published literature also highlighted the positive potential of ‘neighbourliness’ or ‘conviviality’, that is the positive, relatively fleeting, encounters between people in public spaces that could contribute to a feeling of being ‘at home’:190

Strangers sometimes can be positive social connectives. I can walk down the street and I can talk for England and I would just meet someone and speak to them as though I’ve known them for years, and just listening to other people and people having a different outlook in life, that can be a positive thing that can uplift you, so it’s a positive connective.

CP older persons group

Some interventions appeared to have successfully drawn on this reservoir of goodwill and desire to connect. For example, Refuweegee118 has managed to mobilise very large numbers of people in Glasgow to be part of a welcome initiative for newly arrived refugees.

Confidence in encountering such neighbourliness could, however, be very context specific. Unfamiliar places can make people wary, and undermine the potential for conviviality, as can rapid changes in the make-up of an area:

The university is a safe zone. If I’m within the structure of the university, I can go anywhere. When I’m in the wider city, I don’t really fit in. Going for a national insurance number, the vibe was ‘what is going on?’. There should be a way to facilitate this with the job centre as a student. Nobody is facilitating you.

CP student group

There are big issues with neighbours. It’s changing. People move on. Someone has come next door to me, with a different language. A large influx of Nigerian, Polish, Latvian and other, makes me feel less belonging.

CP older people

Community assets and geographies; natural spaces of encounter and connection

The presence, or absence, of spaces and places for social encounters and connections was highlighted as important.

There were concerns about restricted opportunities for everyday interactions in some localities, such as residential buildings not being conducive to sustained and repeated exposure to neighbours, and missed opportunities during new planning projects to create associational spaces that are open to everyone. Although the notion of self-segregating communities has been challenged, some studies did note the limited opportunities in some parts of the country, particularly for children, to interact routinely across ethnic and religious difference. 110,127,130,131

Consultation panel members talked about libraries as important spaces where ‘all sorts of people’ can come together.

Churches and other places of worship were also felt to be important for both established ethnic minority groups and new arrivals as places to feel safe and to connect around a shared religious identity, and, in some cases, a shared ethnic (or national) identity too:

How I met my Jamaicans here in Sheffield is that one Jamaican led to another who led me to another . . . this helped me because she led me to a church and I meet a lot of Jamaicans; the Windrush generation and all that.

CP student group

Sports clubs were also identified as having the potential to bring people together and facilitate new social ties, in some cases across ethnic difference, and in others within self-identifying ethnic or national groups:

I work for [name of place] ethnic minority sports club and we have got 100 women playing sports every week. We’ve got lots of women coming every Saturday from the Tamil community and playing netball. So that kind of sports element brings them together, groups of women.

CP working-age group

However, as noted in Chapter 6, the published evidence on sports activities as sites of positive encounter is mixed, and suggests the importance of facilitation.

Some ethnic communities were identified as lacking permanent community spaces which they could occupy and make their own. For instance, use of a general community venue by Lychee Red, a support group for Chinese older people in Leeds, was felt to be inadequate by members. 267

As well as physical meeting places, some CP members identified virtual communities and networks that could provide an important social support and companionship:

It’s both. So it’s professional in that, so being a woman and being black, in a professional role you’re not just a minority, you’re a super-minority. . . . [this contact makes] you see yourself or your potential self and it’s really inspiring. Other professional elements would be, like, workplace conflicts; I’ve had this myself, where people say things that are, like, really racist or really sexist and you don’t necessarily have the tools to [address this] . . . people know where you’re coming from. On a personal level, it’s just a nice community for discussion and debate.

CP student group; talking about an online network

Importantly, however, these ‘bottom-up’ places of naturalistic encounter could exclude, as well as include, people. An example was given of a local authority that had intervened to ensure that a local market remained a welcoming space for everyone following the development of interethnic group tensions and fear among some people around accessing this facility. Some CP student members shared their experiences of not quite fitting into any of the on-campus ‘community assets’ that were available:

You have that situation where you feel like you don’t fit in anywhere.

I think it works both ways . . . In groups, talk can be ‘British’ in one group and ‘Nigerian’ in another. May exclude. Too British to be Nigerian and too Nigerian to be British: I don’t belong anywhere!

CP student group

Transport

Consultation panel members noted that public transport can be unreliable, and unwelcoming, making people less inclined to venture out of their home. Getting about can be made harder for those who have poor English-language skills, and those who do not drive or have no access to a car. Similar evidence came from some of the published papers:

You find people who don’t welcome you and they are really racist and, you know, they abuse you on the bus and, you know, say words. I been, it happens to me, and one time I was really, really upset and I was crying when I went home.

Female participant, Strang et al. 187

Reproduced with permission from the Scottish Refugee Council

Inadequate transport options also hampered people’s ability to participate in some initiatives,164 whereas other initiatives included special provision to facilitate access. For instance, a group-based psychotherapy intervention for Pakistani women provided taxis with a female companion. 268 More generally, intervention papers talked about using accessible venues, although this could relate to the identity and familiarity of venues as well as their physical location.

Alienating systems and processes

All CPs identified ways in which official systems and processes frequently act to ‘other’ and undermine a sense of belonging among migrants and ethnic minority people. A student CP member described experiences of having to deal with visa arrangements, ‘reminding me that I am different, not a member of society, do not belong here.’ Another CP member described the insensitive processes of the welfare benefit office:

On Fridays, they get Muslim men to come and sign when it is Friday prayer times, exactly, spot-on, that time of the day. These kinds of things, they do get to people.

CP working-age group

And another CP member described poor experiences in a care home:

It’s lack of cultural awareness [affecting my wife’s end-of-life care]. Different religions . . . [My wife] was at the care home, every day they would send an Anglican priest. She’s a Hindu. After a couple of weeks, I said ‘you are out of order, my wife’s a Hindu, you sending a Christian’. After that they identified a Hindu priest . . . They are not sensitive to the cultural needs of different people. It’s a ‘one size fits all’ so you provide one service to the majority community and the rest have to fall in line. Social services have this problem.

CP older persons group

The negative impact of systems and processes on the self-worth and social connectedness was most significant among those seeking asylum and refugees. The system created demands on their time and sapped energy for building social relationships and engaging in relevant activities. 21,187,239 In addition, unstable housing undermined people’s ability and motivation to establish social connections; adding to feelings of uncertainty and insecurity:

And my first experience at Heathrow when I ran and I went to the immigration office and I said ‘I’m in trouble, I’m seeking protection’. It started from there, the way you are treated is like you are a criminal. You lose total control and you are less human, you are an underclass. And the housing is the same, it compounds, so it’s not just having no choice. You get housed in areas where the housing stock is really cheap, poor housing, no aspiration, and you are stuck – no opportunities as well.

CP working-age group

Strang et al. 187 found that access to English-language classes (that could support the social connectedness of asylum seekers and refugees) was impeded by the need to attend appointments, particularly immediately after being granted status.

Consultation panel members also noted a reluctance to complain about organisational practices among migrant and ethnic minority people, for fear of repercussions:

It’s hard to complain when you’re older. If you feel unhappy about the service – feel unable to challenge poor care.

CP older persons group

Staffing: skills, identities and behaviours

Although overt racist discrimination was less commonly identified within organisational settings than out and about at community level, evidence across our sources indicates that the experience of microaggressions is commonplace.

Examples were given of individuals in authority or service-providing positions, through their comments and behaviours, undermining the self-worth of migrant and ethnic minority individuals and contribute to a sense of not mattering, not belonging and not fitting in:106,273,274

You know like they’ve got to eat restaurant-style and we have to eat with our right hand because the prophet – peace be upon him – ate with his right hand, and that’s how we eat. But the schools are telling the kids that they can’t do that, and that’s confusing for the kids.

CP working-age group

My mum was in a home and you notice others who don’t have family or carers. And some carers are not carers, if you know what I mean. Some dread the carer, they want someone who speaks their language and has respect. Some are crying out for help because of loneliness. For the aged, hygiene and incontinency are the big things – they all contribute to loneliness.

CP older persons group

Wilkins and Lall275 found that ethnic minority student teachers experienced stereotypical attitudes among their white peers and that, as teachers, their contribution was often narrowly constructed in terms of their ethnic identity, leaving them feeling isolated and marginalised. Garner et al. 274 found that poor communication between health-care professionals and West African mothers led to difficulties in understanding needs, and an increased sense of isolation among the women.

In one case, past experiences at a local Children’s Centre had been so negative that a support group project had moved to another more neutral venue to encourage attendance and create a positive atmosphere. 245

Again, there was consistent evidence to suggest that people seeking asylum and refugees are perceived particularly negatively, and frequently experience microaggressions:

When they assume I’m an international student, here just to study, the reaction, the conversation that you have is totally different. People accept you, you are international, they respect you. Many times . . . I have said I’m still a refugee, it’s totally different – the reaction. The judgement comes in, you – you are less important, you are not capable. That’s the perception.

CP working-age group

Furthermore, because these individuals typically have few family or friends in the country, insensitive treatment by professionals was felt to be particularly detrimental:

If they do not get the appropriate support from services or even people, it actually knocks their confidence back and then obviously they are not able to actually connect.

CP working-age group

In relation to interventions aimed at tackling isolation and loneliness among migrant and ethnic minority people, as mentioned in Chapter 6, some aimed to adopt explicitly culturally competent models.

A further relevant dimension was the creation of volunteering opportunities in initiatives for current or former recipients. This approach could be helpful in diverse ways, providing a wider range of visible ethnic diversity and language skills among providers, as well as opportunities for volunteers to develop skills and self-confidence. Volunteering provided a much-needed opportunity to demonstrate value and reciprocate, key factors in boosting self-worth:176

Opportunities to help others makes you feel better.

CP older persons group

I feel so empowered to see how far I’ve come. Not just to be now a service user, it was a little bit of, like, upgrade in life, I felt like.

Black African mentor mother, McLeish and Redshaw218

Having volunteers who share the ethnic or migrant identity of potential beneficiaries can also help to develop trust and increase access. McLeish and Redshaw96 noted the challenges to building trust among pregnant women, with asylum-seeking women being fearful and suspicious. In that study,96 volunteer peer supporters reported that they find it easier to be open with other ‘asylum people’ and that being non-judgemental helped in forming a close connection.

However, Vickers176 suggests the need to think carefully about who gets to volunteer when working with refugee (and migrant) groups, as the need for confidence, education and language skills means that opportunities go to people from relatively privileged backgrounds, with the potential to create divisions between these people and refugees who arrive with less human and social capital.

Consultation panel members also raised the issue of negative experiences in workplaces that can undermine self-worth and create a sense of not belonging. In some cases, these experiences were felt to ‘spill over’, generating more chronic loneliness for some people, particularly those for whom their career is an important part of self-identity. The review of published literature did not identify any initiatives based in workplaces, suggesting an important evidence gap:

In the office that I manage, people ask me for the manager, ‘where is the manager?’ and that keeps happening. It’s like ‘how can this African guy be the manager or the in-charge?!’.

CP working-age group

Dominant UK policy and media narratives

A strong message from CP members was that the wider hostile environment towards migrants (and, by extension, ethnic minority people) seriously undermines people’s sense of belonging, and can mean that people do not take up the services and support that are on offer.

Connections to professional staff who might provide support and reduce isolation are compromised because migrants are fearful and find it hard to trust professionals.

Encounters at community level are also undermined. For instance, young Muslims in one study133 reported that, despite their efforts in social action, charity-giving and interfaith involvement, they remain victims of negative media stories. This study also illustrated the way in which ‘bridge-building’ projects, aiming to create positive ties between people of different faiths and ethnicities, are susceptible to being incorporated into popular and dominant narratives around terrorism and security. 133

Ripple effects of ‘remote’ events

This societal backdrop also means that remote events are felt close to home, and directly affect migrant and ethnic minority people’s sense of security and belonging. 127,130

Raw127 examined a school-linking project around the time of the 2005 terrorist attacks and found that primary school children were directly affected. Muslim children reported feeling scared (‘I don’t know if they might be angry with us’), and white British children expressed mistrust (‘suspicious of brown people now’). 127

Targeting and tailoring

Recognising the multifaceted nature of loneliness and its underlying causes, the importance of targeting and tailoring initiatives was highlighted. Enabling people to identify their own needs and select from a menu of support options was suggested. Recognising that people may have multiple needs was also identified as important by CP members and practitioner workshop participants. Published research provided supporting evidence. For instance, Strang et al. 187 noted that refugees report high levels of social connections with friends and family, and yet still express a strong sense of homesickness and loss at being apart from close family members, suggesting that certain needs were unmet. In Chapter 6, we presented evidence that tended to support the contention that interventions that target underlying causes of loneliness are more effective, on average, than those that do not, whether these be carefully tailored approaches or approaches that are multifaceted, allowing a range of potential underlying causes to be addressed.

A further important dimension was the importance of working with, rather than against, people’s existing social ties and connections. The importance of recognising the ways in which interventions can inadvertently undermine existing ties was noted, as well as the need to support people to protect, rekindle and strengthen ties when they can. This was a theme found across our consultation work and the published literature, although many intervention designs were not explicit in terms of these considerations.

Signposting and referring

Linked to targeting and tailoring is the importance of signposting people to appropriate provision.

Concerns were expressed in published literature and CP discussions that, although initiatives may be beneficial to those who are involved in them, those most in need of support often remain unreached (see Chapter 7). In addition, there was evidence that fixed-term interventions can be problematic if they do not include effective ways of linking beneficiaries to more sustainable social networks and sources of support. 96,164,218

There were some examples in the published literature of good partnership-working that supported the identification and signposting of migrant and/or ethnic minority individuals in need of support. 104,190,192

However, a common theme at our practitioner workshop was that professionals, as well as community members, had inadequate knowledge of available resources and services in their local area that could meet the needs of migrant and ethnic minority people. There was a need for better communication across organisations. Furthermore, some people felt that signposting and partnership-working had deteriorated in recent years as funding had been cut and third-sector organisations increasingly felt that they were in competition with each other. White British practitioners, who were in the majority at our workshop despite its focus, also felt ill-informed about community assets that could be relevant to minority individuals, further underscoring the importance of improved links to community-based organisations that serve these populations.

Embedding and diversifying

A further theme was the importance of recognising that people have very varied preferences, and that there is therefore a need for varied provision. CP members told us that not everyone will want to join a ‘group’ and ‘not everyone wants to do gardening!’. It was also pointed out that lots of existing activities and community assets are already protecting against loneliness, such as churches and other places of worship, but that these are not necessarily well supported financially or recognised for their contribution.

Furthermore, a wider range of activities that could have the potential to make an impact are currently overlooked when the issue of loneliness is under consideration. CP members and workshop participants felt that there was work to be done to ensure that public spaces and generic activities (e.g. leisure spaces, libraries) were more conscious of promoting positive social connectedness in general, and more inclusive and welcoming to migrants and ethnic minority people in particular. Positive examples were given as possible role models.

Engaging local people

As noted in Chapter 7, ‘neighbourliness’ appears to play an important role in boosting people’s sense of belonging. CP members, and some published papers, provided examples of how local people had been effectively engaged in initiatives aimed at welcoming newcomers,118,253 as well as some intercultural encounters involving diverse, established communities. School and university students have also been engaged in some relevant projects, although we found no evaluations in the UK literature. These examples raise the possibility of untapped potential for addressing loneliness among our population groups of interest.

Challenging aversion to ethnic-specific provision

A further theme that arose across the data sources was the aversion in current UK policy at both local and national levels to investing in initiatives that are seen as serving only one particular community. There is an important tension between the demonstrated need for opportunities to build and celebrate intraethnic ‘bonding’ ties, and policy orientations that consider such ties to be threatening, and seek instead to prioritise bridging ties across difference. 190

Most participants in our practitioner workshop felt that they were currently expected to design initiatives that are ‘for everyone’. This policy stance did, however, appear to vary somewhat across localities. For instance, the evaluation by Wigfield and Alden267 suggested that, in Leeds, several ethnic-specific initiatives had been supported, and well received, in recent years as part of the Time to Shine programme.

Costs, resources and funding regimes

Very little information was presented on costs of interventions in the published UK evaluations, and there were no assessments of cost-effectiveness.

A large number of the reported initiatives involved volunteers in their provision, suggesting that such initiatives may represent good value for money. On the other hand, in a few cases, authors reported that initiatives were relatively expensive compared with provision for other groups because of the need for ‘culturally sensitive’ additions. 268

Consultation panel members and participants in the practitioner workshop bemoaned the commonly short-term funding regimes that were currently in place, making it difficult to sustain provision, and the small amounts of money that were devoted to tackling complex individual and neighbourhood issues:

In the 80s, there were about 20 luncheon clubs. We are not holding luncheon clubs anymore. It’s cut, cut, cut. The elderly will end up in hospital then, costing more money . . .

CP older persons group

There was evidence in the grey literature too that initiatives commonly involved small investments and were not sustained. 245 For instance, in the intercultural pilots evaluated by Wilson and Zipfel,121 areas were provided with just £7000 to cover project management and £3000 for actual activities in the neighbourhood. Practitioner workshop participants reported that levels of investment in community-level activities were low, and expectations of rapid improvements unrealistic.

Searching and literature retrieved

We adopted a broad and inclusive approach to searching and screening relevant literature. Nevertheless, there was a need for a manageable approach. We opted to restrict our searching to a smaller number of electronic databases than originally planned in the protocol, as the initial yield from three databases was high. In keeping with the theory-informed approach, we decided that an iterative approach using citation follow-up and targeted searches was a more appropriate use of available resources. It is, of course, possible that this design decision resulted in some relevant evidence being overlooked.

In addition, we opted not to include search terms related to religion or religious identity, but, instead, to focus on those related to ethnicity, race and migration. This may have resulted in some relevant material not being captured. CP discussions highlighted the importance of faith to many ethnic minority people.

Furthermore, there may be material residing in other literatures (e.g. relating to health promotion, how to tackle racism, build social capital and community assets, and enhance cultural competence of organisations) that is not explicitly linked through to our proximate determinants or outcomes, and so was not picked up through our searches.

Our assessment of gaps in current UK provision is likely to be compromised by under-reporting of activity in grey and published literature. This may be a particular problem for short-term initiatives targeted at our population groups of interest. We suspect many missed opportunities to document initiatives and to learn about intervention processes and outcomes.

Consultation panels

We believe that the CPs were successful; feedback from participants was overwhelmingly positive. CP involvement importantly shaped the research findings. Nevertheless, we acknowledge that earlier engagement, and a greater depth of participation, would probably have been even more valuable. For instance, earlier engagement might have led us to include religious search terms in our database searches. Employing Cook et al. ’s50 framework, we achieved variable levels of participation across the project life cycle, including ‘compliance’, ‘consultation’ and ‘co-operation’. Importantly, however, we also noted the emergence of significant new understanding (characteristic of Cook et al. ’s50 ‘co-learning’ participation level), although this was not within a framework of action-planning in the current project. Failure to establish a London CP, as originally intended, was also a limitation, as this might have offered complementary insights to those from the panels convened in Leicester and Sheffield.

Conceptualising loneliness

Study findings support the complex nature of loneliness and the importance of recognising its multiple dimensions: emotional loneliness (lack of intimacy); social loneliness, (lack of) sense of belonging, feeling isolated; and feeling unsupported.

Study findings demonstrate the utility of the four proximate determinants model to thinking about the routes to reducing or protecting against loneliness: increase positive social ties and interactions, reduce negative social ties and interactions, increase self-worth, and enhance positive appraisal of existing social ties and interactions:

• Those designing or commissioning loneliness strategies and interventions for local populations may find the model useful in understanding the aetiology of loneliness and potential avenues for action.

• Practitioners seeking to prevent or reduce risk of loneliness among individuals could use the model in identifying underlying causes/risks and tailoring solutions.

Understanding interventions

Study findings demonstrate the value of identifying the functions of interventions, and their intended pathways of effect, rather than focusing narrowly on the superficial form that they take.

Logic models developed for befriending and SSG interventions were well supported by the evidence. Several of the causal chains identified were also demonstrated in other intervention types:

• Those designing or commissioning interventions may find these models helpful in assessing the probable success of proposed or existing initiatives by determining whether or not functions and causal pathways are achieved.

Study findings suggest that interventions that target the proximate determinants of loneliness are more effective than those that do not:

• Those designing or commissioning interventions may find it useful to assess whether or not interventions are likely to be successful in targeting the proximate determinants via tailoring to the particular identified risk factors of a group of beneficiaries; being flexible enough to enable attention to individual risk factors; or involving multiple components that link to all, or most, of the proximate determinants.

Study findings support the assertion that interventions that explicitly aim to boost self-worth are more effective than those that do not:

• Those designing or commissioning interventions may find it useful to assess whether or not interventions integrate components that directly enhance self-worth (e.g. opportunities to reciprocate, to demonstrate competence, to build confidence via new skills) and/or enable user involvement and a sense of ownership, while avoiding overburdening users.

Findings demonstrate that experiencing negative social ties and interactions is common among migrant and ethnic minority people. There was also some evidence that interventions that fail to address this proximate determinant are ineffective at reducing loneliness:

• Those designing or commissioning interventions can assess whether or not interventions are adequately combating negative social ties and interactions when they contribute to loneliness among recipients.

Enhancing existing approaches

Wider system factors and opportunities for action

The study co-produced with practitioners and CP members a ‘loneliness-proofing’ diagram (see Figure 6) that highlights meso- and macro-level processes that influence the risk of loneliness:

• Those designing or commissioning loneliness strategies and interventions for local populations may find the diagram useful for mapping current activity, identifying gaps and highlighting opportunities for more impactful action.

Findings suggested a number of areas in which current UK activity appears to be low:

• Those designing or commissioning loneliness strategies and interventions for local populations may consider whether or not action is needed to reduce exposure to negative social interactions (e.g. provision of safe spaces); create associational spaces where people can mingle free of charge; engender neighbourliness among multiethnic communities; address the lack of material resources and poor access to transport that can undermine social connections for some; equip people with skills to improve social relationships; provide training in, and access to, digital technologies to connect across distance; eliminate microaggressions from significant others (e.g. teachers, bus drivers and service providers); provide information, skills and navigational support for people new to their social context; and improve collaboration, signposting and referral between services and resources.

• Local policy-makers may also find it helpful to consider whether or not broader policy-framing and culture creates (or undermines) an enabling environment in which to tackle loneliness among migrant and ethnic minority groups, and whether or not the contributions of core services and anchor institutions are clear.