To synthesise the literature applying social capital theory or SNA to understand relational influences on help-seeking for MH and SU among people experiencing social marginalisation.

1. How have SNA methods been applied to understand the relational influences on help-seeking for MH and SU among people experiencing social marginalisation and what methodological insights have been gained?

We pre-published our protocol on OSF (www.osf.io/nfkmq).

This review is part of a larger study informed by a Patient and Public Involvement Group consisting of people with experience of justice-involvement. They have confirmed that understanding how to increase early help-seeking for MH and SU is important and will advise on the dissemination of review findings alongside wider study activities.

Eligible papers examined MH or SU help-seeking from any source among adults experiencing social marginalisation beyond sociodemographic characteristics, and applied either social capital theory or SNA. We did not limit context or date, but restricted language to English and our search to peer-reviewed sources.

We sought papers which focused on or discussed help-seeking for MH or SU among adults (18+) experiencing social marginalisation. We regarded help-seeking as an individual’s action(s) to try to elicit, or increase their levels of, help. Papers focused on people already receiving help without searching for additional help were excluded as these were not about seeking help.

We considered aspects of social marginalisation that went beyond sociodemographic factors. This included involvement in the criminal justice system (justice involvement), homelessness, pre-existing mental ill-health, problematic use of drugs or alcohol or gambling.

We included papers if they used social capital theory or applied SNA. While there are wider ways of thinking about relational influences on help-seeking, we focused on literature applying social capital theory or SNA approaches due to their growing application across fields of research, and to keep the review firmly based around relational influences.

Initial scoping searches identified a very limited literature applying social capital or SNA methods to MH or SU help-seeking when excluding papers where the marginalisation experience was solely related to MH or SU problems. Papers where marginalisation was solely related to MH or SU problems offered relevant findings and were included due to the prevalence of MH and SU problems among those experiencing marginalisation due to homelessness or justice involvement.9 16

We searched four databases to identify literature across health, social science, criminology and criminal justice: Criminal Justice Abstracts (CJA), EMBASE, SocINDEX and Web of Science (all databases).

Initial searches for SNA or social capital theory and MH or SU in Web of Science produced a manageable number of publications for review. We thus proceeded in our full search without using search terms for help-seeking or social marginalisation, allowing us to take a nuanced approach to inclusion that accounted for variation in definitions and terminology. As a scoping review, we did not include an exhaustive range of types of social marginalisation, selecting problem SU (alcohol and drugs), problem gambling, mental ill health, homelessness and justice involvement due to their association with multiple disadvantage.9 16

The review team considered papers carefully for whether the phenomena studied could be interpreted as help-seeking for MH or SU. We determined this to be where a person approached another/others to discuss or disclose MH or SU concerns or where someone sought advice, support or help related to their MH or SU. We attempted to capture the inverse, where people were ‘not help-seeking’ by disengaging from or avoiding treatment and support. Consistent with scoping review methods, we did not use every possible term for MH and SU, but selected relevant terms in conversation with a subject librarian that would provide broad coverage.

In each database, we combined search terms for key concepts using Boolean operators. We searched title, abstract, keyword and, where available, subject headings/indexing. In Web of Science, we searched by ‘Topic’ (title, abstract and keywords). Table 1 shows the template search strategy, with the full search strategies included in the online supplemental file 1.

We set up email alerts in Web of Science (which indexed all the included papers and had the largest number of results) to identify new publications as the review progressed until June 2024, reviewed the reference lists of included papers and met regularly to determine if additional searches were needed, although this was ultimately not required.

CC conducted the searches on 26–27th June 2023, imported results into reference management software39 to remove duplicates and imported de-duplicated results into Rayyan40 to facilitate team review. Email alerts for papers meeting our search criteria were checked and reviewed until June 2024 with the intention of adding further papers as they emerged.

In an initial calibration process, three reviewers (CC, DG, RK) independently screened the same 10% of citations at title and abstract level and discussed results to agree a consistent approach. Following this, the three reviewers screened all papers at title and abstract level with at least two reviewers screening each record. We met at the mid- and endpoint of title and abstract screening to discuss inclusions, exclusions and any areas for additional searches. Disagreements were resolved in a three-way discussion, producing a list of papers relevant for full text review.

The three reviewers independently read 10% of papers at full text to determine eligibility for inclusion as an initial calibration process. Following this, all papers were independently screened by at least two reviewers. We again met at the mid- and endpoint to discuss decisions and resolve any conflicts. We intended this stage to be iterative,38 but did not identify a need for further searches. We checked all references lists of included papers and relevant papers that were outside our scope.

Three reviewers (CC, DG, RK) piloted a data charting template before meeting to review the consistency in content and detail of the extracted data. Minor amendments were made to clarify whether a paper’s focus on help-seeking and social capital were explicit or incidental. We included papers that had incidental findings of interest but were not primarily focused on help-seeking/social capital/SNA due to the limited literature. We added extraction of sample size and study aim(s) and streamlined our recording of individual study results. Reviewers divided the remaining papers for data charting, with CC checking all extracts. It was not necessary to contact authors for further details. For each paper, we recorded the data detailed in table 2.

We anticipated we may find more theoretical papers discussing social capital, and because of our methodological focus on SNA, we elected to appraise the quality only of empirical SNA papers. As there are no validated critical appraisal tools for SNA, at least one reviewer appraised the quality of each paper using the Mixed Methods Appraisal Tool (MMAT), selected for its ability to support efficient appraisal of various study types.41 We were cognisant of potential limitations in elucidating methodological strengths and limitations of SNA designs and in this respect were testing the utility of MMAT for SNA.

We synthesised the results using data summaries and visualisations, before developing a narrative synthesis addressing each question. For papers discussing social capital, we summarised, grouped and compared these according to the number and type of papers, publication dates and rate, and sample, sample demographics, country, focus on MH or SU (or both), theorists and types of social capital. For SNA papers, we summarised the number and type of study, sample, sample demographics, country, focus on MH or SU (or both), and methods applied. These were discussed as a wider team at the midpoint to capture impressions and inform the final synthesis and reporting.

We identified and screened 4487 papers at title and abstract level and reviewed 122 at full text. We included 27 papers. Five were studies focused on help-seeking, of which four were identified in database searches and one from wider reading. The latter had not used ‘social network analysis’ as a term in full, thus not being identified in initial searches (see figure 1). An additional 22 papers were retained in the review due to incidental references to help-seeking. While help-seeking was not the focus, they provided useful synthesis material to better understand the phenomenon.

We summarise the characteristics of the included papers here and present our full data extraction in online supplemental table S1 (Social Capital Papers), and online supplemental table S2 (Social Network Analysis Papers).

Applying the MMAT41 to SNA papers provided an initial framework for assessing quality. While this allowed us to identify broad areas of strength and limitations, it provided little direction to the specific considerations for different types of social network studies. The main areas where we identified limitations were in clarity of reporting methods, particularly in how qualitative findings were synthesised and how multiple methods were integrated in a single study to reach conclusions. See online supplemental table S2.

Twenty-three papers referred to social capital. Only two studies were focused on help-seeking,53 56 of which only Perry and Pescosolido56 directly applied social capital in their work. The remaining 21 were focused on other phenomena but with incidentally relevant findings.

Most papers were qualitative, with two literature reviews, 46 47one mixed methods study50 and one quantitative study.56 Eighteen used or described social capital, while five only made a short reference to it.52 54 58 59 66

There is limited literature that uses social capital theory or SNA to examine relational influences on help-seeking for MH and SU among people experiencing social marginalisation. We identified few studies that sampled people experiencing homelessness,55 57 64 and the only paper discussing justice-involved people was a literature review.47 In the few studies explicitly focused on help-seeking, the term ‘help-seeking’ was not widely used. This may have implications for future reviews attempting to examine help-seeking for MH, SU or other needs, potentially resulting in few results. A priori theorisation and decisions about what the review team considers ‘help-seeking’ will be beneficial.

High bonding capital among people experiencing social marginalisation to others within minority communities, in the absence of bridging and linking capital, could deter help-seeking for MH and SU. This extends the evidence on the health risks of high bonding capital32 34 by highlighting that deterrence of help-seeking may be a mechanism by which social capital affects health. We noted two ways in which this mechanism may operate. First, help-seeking may present a risk of losing social capital in networks and communities where norms and culture do not support engaging with authoritative structures, or where there is network and community stigma towards people who have MH or SU problems, thus deterring help-seeking. Perceived stigma towards people in need of help, translating into internalised stigma, has been shown to negatively influence MH help-seeking.73 While the risk of losing social capital also applies to people who possess high levels of bridging and linking capital, arguably the risk is more pronounced for people who rely on being part of a close-knit community. Second, an absence of linking and bridging capital can reduce access to varied information74 about when help can and should be sought, what is available and how this might be experienced, leaving people without the information needed to decide to seek help.

Viewing social capital as the sum of resources within a network, having lower bridging and linking social capital could indicate a network holds fewer resources, and thus influences people to make deliberate decisions about from whom to seek help. There were references to both seeking out friends and family and avoidance of doing so, where people were conscious of not burdening others with limited resources. This is in contrast to evidence that people make greater use of weak ties and avoid stronger ties when disclosing sensitive personal circumstances.75 However, Small’s conclusions were developed from studies with people with a greater degree of social integration (university students), who may have greater access to weak ties through their social positionality, and did not examine highly stigmatised issues. Findings from our review indicate that relationship qualities, including closeness, may be particularly important for sensitive and highly stigmatised illnesses among people experiencing social marginalisation. The small volume of evidence identified in this review limits our ability to draw firm conclusions, and the importance of tie strength may warrant greater exploration in populations where there are low levels of bridging capital.

Selectivity about from whom help is sought was clear, with preferences about alters’ social roles varying between studies.49 56 64 However, relationship qualities were more important than alters’ respective roles. Specifically, those relationships characterised as caring and perceived as likely to be helpful were the ones that were approached. Further, the alters that participants did or would approach changed over time, reflecting the influence of past experience on where, when and from whom someone may seek help. This contrasts with less deliberative approaches to help-seeking for lower-level concerns among students75 76 and is more consistent with theories that help-seeking is a socially influenced, but still deliberative, action.22 23

Qualitative studies showed that at times, participants characterised professionals as friends, highlighting a need for a clear understanding of how people define the roles of alters in their networks, especially when measuring this in quantitative approaches. In the quantitative analyses, variables associated with help-seeking were identified, with some conflicting results. More studies that account for different experiences of marginalisation in different contexts but using the same variable operationalisation could progress understanding of the relational influences on help-seeking and the best ways to apply social capital and SNA to study these.

Trust is debated by social capital theorists regarding whether it is characteristic of relationships or a more generalised community-level asset and whether it contributes to, results from or is integral to social capital.77–80 Trust is an important relationship quality that enables access to social capital, and network levels of trust in practitioners were associated with help-seeking. Mistrust of authoritative services is a well-elucidated cultural norm within some communities that deters help-seeking, for example, among people in prison18 and women who use drugs.81 Trust in healthcare providers at an individual level varies across communities, partly explained by levels of community trust and community social capital.82 The importance of network (and community) culture and attitudes towards different sources of support has received little examination in groups who experience marginalisation by some communities/social groups, but who have strong bonding capital in communities sharing their marginalisation experiences.44

Structural factors were considered to a limited extent (eg, law and policy that fails to allow for harm reduction opportunities, sociodemographic characteristics). Future research that elaborates exactly how these macro factors influence help-seeking for MH and SU among people experiencing social marginalisation is required, particularly accounting for the intersecting experiences that further compound disadvantage.83 Better understanding structural influences can inform and enhance the impact of ‘Inclusion Health’ policies and practice84 to assist those in greatest need of MH and/or SU support.

SNA methods present a useful means of empirically examining help-seeking among people experiencing social marginalisation, through use of visual network mapping, the application of multilevel models and its potential for informing intervention development. Although SNA in health and social care research is increasing, the limitations in the wider evidence base are evident here, including the need to better understand relational influences on health-related behaviour, the mechanisms by which social networks have their effects and greater use of longitudinal data to understand change over time.85

No studies used the term help-seeking specifically. For relevant papers to be identified for future reviews, authors of primary research may consider adopting consistent terminology or including help-seeking as a keyword.

We only quality appraised SNA studies and did not exclude papers on quality grounds. Our main observation was limitations in the clarity of reporting, particularly in how qualitative SNA findings were synthesised and where multiple methods were combined in a single paper. We included 27 papers and provide a broad overview of how social capital and SNA are applied; thus, our findings can be read with sufficient confidence for that purpose. Our synthesis of findings from the social capital studies should be interpreted with awareness that these papers were not quality appraised.

Most papers included participants who used drugs, and several included participants experiencing MH challenges. Although homelessness, sex work, justice-involvement and poverty were all represented, and there are significant overlaps between populations, there is clearly a need for research that examines the unique relational dynamics influencing help-seeking when people experience social marginalisation due to different and multiple experiences. Core demographics were not or could not be reported for some studies. Reporting precise sample demographics and characteristics is essential for future reviews that seek to take account of the intersecting experiences of social marginalisation including intersections between age, sex/gender, ethnicity or sexual orientation.

We conducted the review using rigorous methods, pre-publishing our protocol and adopting PRISMA reporting standards. We searched in English and excluded papers where the full text was unavailable in English. This may have resulted in relevant literature in other languages being overlooked, particularly in non-Western contexts. We nonetheless identified studies from countries with different social and cultural dynamics across five continents.

While we focused on selected marginalisation experiences, there are likely to be implications for other populations, such as those marginalised by poverty or long-term unemployment, from minority ethnic groups, diverse sexual orientations or different gender identities. However, the distinct relational influences on MH and SU help-seeking in each of these groups and others should be examined, in addition to where people have intersecting experiences of marginalisation.

While scoping reviews often omit quality appraisal, we appraised SNA studies. It was possible to use MMAT, but it lacked detail for the interrogation of SNA elements. Development of a bespoke tool that can be used for studies adopting this method may be beneficial.

Not all full texts were screened by two independent reviewers as originally intended. At initial and interim team meetings, we determined that the level of consistency achieved between reviewers was sufficient to allow a single review of remaining papers. Additionally, extraction and synthesis were conducted by a single reviewer. Two reviewers at each stage may have added further rigour to the review.

Social capital is not the only social theory that may be relevant, and SNA is not the only way to study it. Future reviews could assess how other theoretical or methodological approaches have been applied to understand the relational influences on help-seeking for MH and SU among people experiencing social marginalisation.

Not applicable.

Not applicable.