1 INTRODUCTION

Social isolation has been linked to numerous health risks, including increased depression and mortality (Hämmig, 2019; Holt-Lunstad et al., 2015). The risks associated with social isolation have been compared to that of smoking and obesity (Pantell et al., 2013). Social isolation, and many of its associated health risks, disproportionately affect some populations over others, further contributing to health inequities. Parents raising children in low-income, under-resourced communities are at a heightened risk for experiencing social isolation and its effects due to limited access to resources that bridge and nurture their social connections (e.g., accessible transportation, safe neighborhoods) (Bess & Doykos, 2014; Keating-Lefler et al., 2004; Rank et al., 1998). This is particularly concerning as social isolation among parents has been associated with a greater risk of child maltreatment (Gracia & Musitu, 2003) and increased health problems for both parents and their children (Thompson et al., 2020). In acknowledgment of this disparity, the National Academies of Sciences, Engineering, and Medicine has made reducing social isolation a leading health indicator for improving social environments and closing the gaps in health equity by 2030 (National Academies of Science Engineering and Medicine, 2020).

Social connectedness (SC) refers to a person's perception of belongingness and connection to other people, or a community (Haslam et al., 2015; Lee & Robbins, 1995), and has been linked to reduced social isolation and improved physical and mental health outcomes (Cohen, 2004). SC may be a modifiable factor that could potentially improve health outcomes in vulnerable populations through the optimization of interventions that foster a sense of belonging and connectedness (O'Rourke & Sidani, 2017). Although existing research has explored SC in populations such as older adults (Haslam et al., 2015; O'Rourke & Sidani, 2017) and individuals with mental health disorders (Hare Duke et al., 2019), little is known about SC and how it relates to health outcomes among low-income families. Given that children's health and well-being are integrally linked to their parents' health and well-being (National Academies of Sciences Engineering and Medicine, 2016), understanding the role of parents’ SC in supporting health outcomes across two generations is an important area of study.

1.1 SC versus social support

SC and social support are related, yet different constructs. Both focus on the qualities of people's relationships with others and have been associated with mental health outcomes (Ozbay et al., 2007; Wang et al., 2018). However, SC and social support are characterized by different aspects of social relationships. Specifically, social support refers to an exchange of resources (e.g., information or money) and has been characterized by four types of support: informational, emotional, instrumental, and appraisal (Mohd et al., 2019; Taylor, 2011). In contrast, SC refers to an individual's sense of belongingness in their relationships with other individuals and groups. These conceptual differences have important implications for how one might approach interventions designed to improve health outcomes by building social support systems versus building social connections within those systems.

The concept of social support has a significant presence in social science literature and its definition and measurement are well-established (Taylor, 2011). There have been several reviews of the literature on social support and related health outcomes and readers are referred to these sources for more information (Lindsay Smith et al., 2017; Mohd et al., 2019; Wang et al., 2018). In contrast, the concept of SC remains underdeveloped, despite increasing interest from health researchers (Holt-Lunstad et al., 2017; O'Rourke & Sidani, 2017). For example, a 2019 National Institutes of Health funding announcement sought proposals to study mechanisms by which SC influences health, well-being, and recovery from illness and vulnerable populations (US Department of Health and Human Services, 2019). The announcement highlighted the potential health benefits of SC as well as the dearth of research examining how SC is developed and its effects on health across the lifespan or generations. This review will focus specifically on this understudied concept of SC.

The purpose of this integrative review is to synthesize existing literature on SC among parents raising children in low-income communities. Specifically, this review aims to: (a) describe the current state of the science surrounding SC in parents with low incomes, (b) summarize how SC is currently defined and measured, (c) evaluate the quality of the published research, and (d) identify gaps in the literature and recommendations for future research.

2 METHODS 2.1 Literature search strategy

Before conducting the search, a health science librarian was consulted in the development of the search strategy (Supporting Information Appendix 1). The search was conducted in March 2021 using five electronic databases: PubMed, CINAHL, Embase, PsycINFO, and Web of Science.

Search terms were established using a combination of medical subject headings (MeSH) and nonindexed terms. Given that SC is a developing concept, it is not an indexed term. Therefore, additional terms were chosen that reflected “a person's perceived sense of belonging to others or the environment” (Haslam et al., 2015) to maximize our ability to identify relevant articles. The following additional terms were included in the search: (a) social cohesion, the sense of belonging among groups in society and the extent of their connectedness (Manca, 2014); (b) group cohesion, the extent individuals in a cohesive group express their sense of belonging to the group (Dyaram & Kamalanabhan, 2005); and (c) social/group identity, individuals' perception of who they are based on groups in which they feel a sense of belonging to (Hogg et al., 2017). The phrase “sense of belonging” was also included to identify potential articles that discuss this concept without using an explicit term. As noted above, social support was excluded from the search to distinguish literature on exchangeable resources (e.g., information) gained through social relationships from a sense of belonging and connection with others.

2.2 Inclusion and exclusion criteria

Studies were included if (a) the population of focus was parents of dependent children with low incomes or living in low-income communities, (b) outcomes in the study were parent-centered (i.e., study outcomes were focused on parents, not just children), (c) it was an empirical study, (d) the concept of SC (or related search term as described above) was a study variable or emerged in qualitative findings, and (e) the study was published in a peer-reviewed journal before March 2021. Studies were excluded if (a) they were not published in English, (b) the sample consisted entirely of participants that did not have children at the time of the study (e.g., participants pregnant with their first child), or (c) the authors used a measure to assess SC that was specifically designed to evaluate a different term excluded from the search (e.g., studies using a social support measure to evaluate SC). Studies that focused on a population or community level of SC (e.g., overall neighborhood level of SC), rather than an individual's perception of SC (e.g., parent's perception of neighborhood SC), were also excluded.

2.3 Screening

Search results were uploaded to Covidence (Veritas Health Innovation, n.d.), a software designed to manage and organize literature for reviews. After removing duplicate results, two authors independently performed a title and abstract screening. Articles included after the initial screening were then independently evaluated in a full-text review by the same two authors. If there was a disagreement in an article for inclusion, a third author read the article and acted as a “tie-breaker” to determine the outcome.

2.4 Data evaluation

The following information was extracted from all studies (1) setting in which the research took place, (2) sample size and participant characteristics, (3) research design and methods, and (4) definition of SC or related term, when available. For quantitative studies, measures of SC and related health outcomes (if applicable) were extracted. For intervention studies, intervention programs, settings, programmatic factors, and major outcomes related to SC were extracted. For qualitative studies, text labeled as “results” in the study reports was considered as data for extraction (Thomas & Harden, 2008). In studies that collected SC data from multiple sources (e.g., parents, providers, children), only data from parents were included.

2.5 Quality assessment

Included articles were assessed for quality using the Johns Hopkins Nursing Evidence-Based Practice (JHNEBP) Evidence Level and Quality Guide (Dang & Dearholt, 2018). JHNEBP is a well-established quality appraisal tool used to evaluate both qualitative and quantitative studies (Dang & Dearholt, 2018). Per the guide, articles were rated for their level of evidence on a scale from Level I (high evidence level, e.g., experimental studies, randomized controlled trials) to Level V (low evidence level, e.g., nonresearch design, case reports). Quality was then assessed using the guide's criteria of study sample size, recommendations, and generalizability of results with “A” indicating high quality, “B” good quality, and “C” low quality (Dang & Dearholt, 2018). Article quality was assessed independently by two authors. Ratings were then discussed to determine inter-rater agreement. When a discrepancy occurred, the quality guide was referenced and a third author acted as a “tie-breaker” until a consensus was reached. The included studies’ evidence and quality ratings are reported in Table 1.

Table 1. Characteristics of included studies on social connectedness and related terms among parents with low-incomes First author (year) Setting Sample (N) characteristics Study methods Guiding theory Intervention Quality appraisal Quantitative studies Acri (2019)

US

Poverty-impacted community

N = 32 caregivers of children aged 7–11

91% female, 61% mothers

52% Black, 44% White, 67% non-Hispanic/Latino, 30% household income <$10k/year

Substudy of National Institute of Mental Health Study

Cross-sectional descriptive

Self-report

None specified 4Rs and 2Ss Level IIIB Adaji (2019)

Nigeria

Rural community

N = 161 mothers in low-income setting

16.3% pregnant for first time

53.5% had 1–4 children

30.2% had >5 children

Prospective observational study

Self-report, group facilitator report, and objective measure

None specified

Prenatal care program based on centering

pregnancy

Level IIIB Booth (2020)

US

Pittsburgh

Urban

N = 185 low-income adolescent males and their parents

Adolescents: 56.44% White, 34.67% Black, 8.89% mixed race

Parents: 89.6%–93.5% mothers; 64.89% White, 34.22% African American, <1% mixed race

Longitudinal descriptive study

Secondary data from the Pitt Child & Mother Project

Self-report

Social disorganized theory N/A Level IIIA Brisson (2012)

US

Boston, Chicago, and San Antonio

Urban

N = 1495 low-income mothers (20% poverty rate or more)

43% Black, 48% Hispanic, 8% White

Longitudinal descriptive study

Secondary data from welfare, children, and families: A Three-cities study

Self-report

None specified N/A Level IIIA Brisson (2019)

US

Western United States

Urban

N = 52 families (37 intervention group, 15 nontreatment group)

96% female, 45% Latino, 22% Black, 14% White in 3 low-income neighborhoods

Quasi-experimental study

Self-report

Ecological system theory

Your family, your neighborhood

(YFYN)

Level IIB McCloskey (2019)

US

20 large cities

Urban

N = 3876 low-income mothers, 48% Black, 25.9% Hispanic, 22.5% White

Cross-sectional

descriptive study

secondary data from Fragile Families and Child Wellbeing Study

Self-report

Pearlin's stress process model N/A Level IIIA McLeigh (2018)

US

South Carolina

N = 483 low-income primary caregivers of children, 68.4% married, 12.8% separated/divorced

70.3% White, 23.6% Black, 3.7% Hispanic

Cross-sectional

descriptive

study

Self-report and objective data

None specified N/A Level IIIA Prendergast (2019)

US

20 large cities

Urban

N =3529 children and their mothers

Longitudinal descriptive study

Secondary data from Fragile Families and Child Wellbeing Study

Self-report

Social ecological model N/A Level IIIA Yuma-Guerrero (2017)

US

California

N = 2750 mothers

31.6% incomes at or below federal poverty line

52.8% Latina, 24.1% White, 6.2% Black

Cross-sectional descriptive study

Secondary data from the Geographic Research on Wellbeing study

Self-report

N/A N/A Level IIIB Qualitative/mixed methods studies Bess (2014)

US

Nashville

Urban

N = 69 parents of Children 0–4 in low-income neighborhoods, 85% Black, 87% women, 81% single, 84% unemployed Qualitative Prilleltensky's model of well-being Tied together Level IIIA Curry (2019)

US

Midwest

Urban

N = 59 parent participants of 12 focus groups at six elementary schools

Majority mothers

Qualitative Social network theory N/A Level IIIA Davison (2013)

US

Urban

N = 89 low-income parents/caregivers of children enrolled in HeadStart

91% female, 52% White, 22% Black, 10% Hispanic

Mixed-methods (quantitative measures self-report) Family ecological model N/A Level IIIC Eastwood (2014)

Australia

Sydney

Urban

N = 8 mothers in low-income community Qualitative None specified N/A Level IIIB Lipman (2010)

Canada

Ontario

Urban

N = 8 Single-mothers, 50% income <$15,000 Qualitative None specified Community-based program, unspecified Level IIIB Parsons (2019)

US

Cincinnati

Urban

N = 20 (15 mothers/grandmothers, 5 neighborhood “block captains”)

65% White, 30% Black, 5% mixed race

Low-income neighborhood residents

Qualitative None specified Healthy homes (HH) Block by block (HH) Level IIIB 2.6 Analysis

The authors synthesized quantitative and qualitative studies separately, then integrated findings from both sources where applicable (Whittemore & Knafl, 2005). Given that the included studies addressed multiple research questions across diverse settings, a narrative synthesis was conducted to examine findings from the quantitative studies. Narrative syntheses are used to compare and summarize quantitative data when statistical comparison across studies is not feasible (Lisy & Porritt, 2016; Popay et al. 2006; Whittemore & Knafl, 2005). Two authors first conducted the narrative synthesis independently, and then three authors met to compare findings and discuss results. The agreed-upon findings were analyzed and are presented below by specific outcomes.

A thematic synthesis was used to synthesize the results of qualitative studies (Thomas & Harden, 2008). This method uses thematic analysis to integrate the findings of multiple qualitative studies (Braun & Clarke, 2006). Three stages of analysis were conducted iteratively starting with independent line-by-line coding of qualitative findings (Thomas & Harden, 2008). Line-by-line coding represents the translation of concepts from one study to another (Britten et al., 2002; Qureshi et al., 2006). Codes were developed when necessary and checked for consistency to see if an additional level of coding was needed (Thomas & Harden, 2008). Two authors then compared codes and checked for inter-coder agreement. The agreed-upon codes were grouped into descriptive themes representing participants’ perspectives on SC. Analytic themes were then developed based on the descriptive themes to address the aims of the review (Britten et al., 2002; Thomas & Harden, 2008).

3 RESULTS

The search resulted in 978 unique records. A total of 951 records were removed through title and abstract screening leaving 27 articles for full-text review. Following full-text review, 12 articles were excluded leaving 15 articles for data extraction (see Figure 1).

Review process

3.1 Study characteristics

Table 1 highlights the key characteristics of the included study. Twelve studies were conducted in the United States, one in Australia, one in Canada, and one in Nigeria. Most participants were mothers. Seven studies included only mothers, six included a range of caregivers (i.e., mothers, fathers, grandparents), and one included mothers and grandmothers. One study did not report specific caregiver roles (McLeigh et al., 2018).

Nine studies were quantitative and six were qualitative in design. Six studies included an intervention. Of the nine quantitative studies, five utilized a cross-sectional design, three used a longitudinal design, and one used a quasi-experimental design. All included studies, except one, were rated as Level III evidence, indicating nonexperimental study designs. The quasi-experimental study by Brisson et al. (2019) was rated as Level II. Seven studies were rated as high-quality, Level A, for demonstrating consistent, generalizable results with adequate study designs and sample sizes. Seven studies were rated as quality Level B, indicating good quality by demonstrating reasonably consistent results for the design and sample size. One study was rated as Level C, poor quality, indicating an inadequate study design for the conclusions made (Dang & Dearholt, 2018). Due to the limited number of studies the search produced, the authors decided to not exclude articles based on low-quality ratings to ensure the most comprehensive review of existing literature.

3.2 SC definition, measures, and guiding theories

Wide variability exists in the terms and measures used to study parental SC in the literature (Table 2). Six studies used the term “social connectedness” or “connectedness” to describe the concept under study. All of these studies used qualitative methods. Ten studies used “social cohesion” to describe parents’ perception of belonging in their relationships. Six of these studies specifically focused on neighborhood social cohesion. One study did not specify an exact term related to SC but reported results in line with the concept of SC (Lipman et al., 2010). The most commonly used measure focused on assessing neighborhood social cohesion; six quantitative articles described using a Likert-scale measure with questions related to parents’ perception of their neighborhood.

Table 2. Measures and definitions of social connectedness and related terms used in studies with parents with low incomes Quantitative studies First author (year) Definition of social connectedness or proxy term Quantitative measures for social connectedness Acri (2019) Cohesiveness: feelings of belonging, understood, and accepted by group members -29-item measure for participants’ perception of the 4Rs 2Ss intervention including 3 items measuring cohesion Adaji (2019)

Cohesion, Group cohesiveness

Definition not specified

-3-item measure of cohesion Booth (2020)

Neighborhood cohesion-

Definition not specified

-5-item, Likert scale measure on neighborhood cohesion Brisson (2012) Social Cohesion: the ability to establish positive relationships and build trust with neighbors -4 items on scale developed to measure social cohesion, informal social control, and collective efficacy Brisson (2019) Neighborhood social cohesion: Definition not specified -4-item, Likert scale measure of neighborhood social cohesion, from the Project on Human Development in Chicago Neighborhoods McCloskey (2019) Neighborhood social cohesion: Mothers’ level of trusting relationships and collective social norms among individuals in a shared community

-Social Cohesion and Trust Scale

-5-item scale, statements of neighborhood social cohesion

McLeigh (2018) Neighborhood social cohesion: Mutual trust and shared expectations among neighbors

-Social Cohesion Scale

-5-item Likert-type scale

Prendergast (2019) Neighborhood social cohesion: mutual trust and support among neighbors -5-item, Likert scale measure on neighborhood social cohesion Yuma-Guerrero (2017) Neighborhood social cohesion: the extent of connectedness and solidarity among residents in a neighborhood -6-indicator, Likert scale measure on social cohesion in their neighborhood Qualitative studies Definition of social connectedness or proxy term Emerging qualitative themes Bess (2014)

Social connections

 Definition not specified

Access to SC

  Scarcity

  Protective avoidance

Connectedness

  Environment of acceptance

  Normalization

Curry (2019)

Social connections

 Definition not specified

Access to SC

  Scarcity

Connectedness

Davison (2013)

Social connectedness

 Definition not specified

Access to SC

  Scarcity

  Protective avoidance

  Loss

Eastwood (2014)

Connectedness

Social cohesion

 Definitions not specified

Access to SC

 Scarcity

  Protective avoidance

Connectedness

Lipman (2010) Social connectedness term not specified: Article discussed mothers’ connections to participants in a community-based support/educational group that included components of social connectedness

Access to SC

  Scarcity

  Loss

Connectedness

  Environment of acceptance

  Normalization

Parsons (2019) Social connection: the ways individuals connect via physical, behavioral, social-cognitive, and emotional pathways

Access to SC

  Scarcity

  Protective avoidance

Connectedness

 Environment of acceptance

 Normalization

Half of the studies provided a theory for guiding their SC research and of those that did include a theory, there was little consistency in the frameworks cited. The most common theory used to frame SC, identified in two studies, was Ecological Theory. Table 1 details the various measures and theories that were used to operationalize the concept of SC.

3.3 Quantitative synthesis

Findings from the nine quantitative studies centered on three overarching themes related to parental SC: parental mental health outcomes, parents' connections to their own community and its resources, and cross-generational outcomes of SC.

3.3.1 Parental mental health outcomes

Parental mental health outcomes refer to outcomes related to any mental health condition, such as anxiety or depression. McCloskey and Pei (2019) reported higher SC (framed as “neighborhood social cohesion”) was related to lower parenting stress. The study also reported lower parenting stress acted as a partial mediator between SC and maternal anxiety and depression (McCloskey & Pei, 2019).

3.3.2 Connection to community

Five quantitative studies reported various SC outcomes related to parents’ connections to their community, that is, parents’ reported perceptions of or interactions with their communities or neighborhoods, and related resources (Acri et al., 2019; Adaji et al., 2019; Brisson, 2012; Brisson et al., 2019; Yuma-Guerrero et al., 2017). Yuma-Guerrero et al. (2017) found SC-mediated relationships between mothers’ perceptions of neighborhood safety and their engagement in physical activity, suggesting that greater SC may be helpful in improving perceptions of neighborhood safety and parents’ engagement in related activities within the neighborhood. Brisson (2012) reported that SC in their communities (described as “neighborhood social cohesion”) was associated with lower levels of food insecurity.

Interventions were also critical in creating opportunities for parents to gain SC within their communities. Acri et al. (2019) found caregivers who participated in the 4Rs 2Ss program, a family group intervention for children with behavioral difficulties, reported high SC with other caregivers within the program (framed as “group cohesion”; Acri et al., 2019). Brisson et al. (2019) reported low-income families participating in the intervention, “Your Family, Your Neighborhood,” designed to improve neighborhood social cohesion, showed improvements in SC (framed as “neighborhood social cohesion”). However, these improvements were not significantly greater than those in a comparison group that did not receive the intervention. Adaji et al. (2019) found SC formed during a prenatal care program in Nigeria to be associated with better program outcomes as mothers who participated in the program had higher levels of SC (framed as “group cohesion”) and increased knowledge in pregnancy issues at the end of the program. These findings suggest group-based interventions may be successful in increasing SC among parents.

3.3.3 Cross-generational outcomes

Three studies reported findings of how parents’ SC (framed as “neighborhood social cohesion”) influences outcomes for both themselves and their children. For example, McLeigh et al. (2018) found that parents’ perceived SC mediated the impact of neighborhood poverty on child abuse, suggesting increased SC may help to decrease rates of child abuse (McLeigh et al., 2018). Similarly, Prendergast and MacPhee (2020) reported increased SC was significantly associated with decreases in mothers’ aggression toward their children in early childhood (Prendergast & MacPhee, 2020). Finally, Booth and Shaw (2020) discussed perceptions of parental SC were positively associated with parental monitoring of adolescent males in the Pitt Mother and Child Project, suggesting increased SC may be helpful in parenting older children (Booth & Shaw, 2020).

3.4 Qualitative synthesis

Two central themes related to low-income parents’ experience with SC were discovered from six qualitative studies: (1) parents’ access to SC and (2) connectedness.

3.4.1 Access to SC The theme of “access to SC” was found across all qualitative studies. This theme characterized parents’ access, or lack of access to SC, contributing to either a sense of belonging or a sense of isolation. Often parents described a lack of access to SC within their community, which contributed to subthemes of scarcity and loss. Scarcity refers to a lack of resources that parents experience in their physical, social, or financial environments. Parents in all qualitative studies described their experiences of scarcity in multiple environments, which hindered their ability to participate in activities or connect with others. Furthermore, many parents described purposely having scarce connections as a form of protective avoidance due to concerns of exposing their children to detrimental social influences in communities they felt unsafe in (Bess & Doykos, 2014; Davison et al., 2013; Eastwood et al., 2014; Parsons et al., 2019). For example, one parent described protective strategies in response to harsh living environments, which in turn created disconnection:

There's a lot of people that's protective over their homes and their children due to…the crime rate. So it's kind of hard … to connect with people. (Parsons et al., 2019, p. 9)

Loss is also salient in parents’ experience with a lack of access to SC (Davison et al., 2013; Lipman et al., 2010). The loss of relationships and associated connections was significant to some who became single parents. For example, Lipman et al. (2010) found single mothers enrolled in an educational/support group attributed their feelings of isolation to a loss of connections after separating from their partner:

‘I found that I felt absolutely alone in the absolute world.’ …many of the women disclosed that their connections to their social circles of friends were severed when their marriages ended. (p. 4)