Advertisement
Browse Subject Areas
?

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

  • Loading metrics

Should social disconnectedness be included in primary-care screening for cardiometabolic disease? A systematic review of the relationship between everyday stress, social connectedness, and allostatic load

  • Anders Larrabee Sonderlund ,

    Contributed equally to this work with: Anders Larrabee Sonderlund, Trine Thilsing, Jens Sondergaard

    Roles Conceptualization, Data curation, Formal analysis, Methodology, Project administration, Writing – original draft, Writing – review & editing

    asonderlund@health.sdu.dk

    Affiliation The Research Unit for General Practice, Department of Public Health, University of Southern Denmark, Copenhagen, Denmark

  • Trine Thilsing ,

    Contributed equally to this work with: Anders Larrabee Sonderlund, Trine Thilsing, Jens Sondergaard

    Roles Conceptualization, Methodology, Supervision, Writing – review & editing

    Affiliation The Research Unit for General Practice, Department of Public Health, University of Southern Denmark, Copenhagen, Denmark

  • Jens Sondergaard

    Contributed equally to this work with: Anders Larrabee Sonderlund, Trine Thilsing, Jens Sondergaard

    Roles Conceptualization, Supervision, Writing – review & editing

    Affiliation The Research Unit for General Practice, Department of Public Health, University of Southern Denmark, Copenhagen, Denmark

Abstract

In the present review, we argue that social disconnectedness could and should be included in primary-care screening protocols for the detection of cardiometabolic disease. Empirical evidence indicates that weak social connectedness represents a serious risk factor for chronic diseases, including cardiovascular disease, diabetes, and various cancers. Weak social connectedness, however, is largely regarded as a second-tier health-risk factor in clinical and research settings. This may be because the mechanisms by which this factor impacts on physical health are poorly understood. Budding research, however, advances the idea that social connectedness buffers against stress-related allostatic load–a known precursor for cardiovascular disease and cancer. The present paper reviews the empirical knowledge on the relationship between everyday stress, social connectedness, and allostatic load. Of 6022 articles retained in the literature search, 20 met predefined inclusion criteria. These studies overwhelmingly support the notion that social connectedness correlates negatively with allostatic load. Several moderators of this relationship were also identified, including gender, social status, and quality of social ties. More research into these factors, however, is warranted to conclusively determine their significance. The current evidence strongly indicates that the more socially connected individuals are, the less likely they are to experience chronic stress and associated allostatic load. The negative association between social connectedness and various chronic diseases can thus, at least partially, be explained by the buffering qualities of social connectedness against allostatic load. We argue that assessing social connectedness in clinical and epidemiological settings may therefore represent a considerable asset in terms of prevention and intervention.

Introduction

Background

Cardiometabolic diseases (CMD) such as cardiovascular disease (CVD), type 2 diabetes, and chronic obstructive pulmonary disease are the leading cause of premature death globally. In 2015, 17.7 million people died from CMD, including most predominantly coronary heart disease (7.4 million) and stroke (6.7 million) [1]. A plethora of empirical studies have shown that the majority of CMDs can be prevented by addressing behavioral risk factors such as unhealthy diet, tobacco and alcohol use, and leading a sedentary lifestyle (collectively the Big 4) [2]. Most research into the prevention of CMD to date has thus focused on changing behavior as it relates to the Big 4. Recent studies, however, indicate that there are other equally important modifiable risk factors beyond the Big 4, which are often overlooked in both research and practice. In particular, studies have emphasized the significance of social disconnectedness as a central determinant of CMD. Social connectedness refers to the actual and perceived nature and extent of the individual’s ties to family and friends, social groups and networks [3]. In a noteworthy metanalysis of 148 studies (approximate collective N = 300.000), Holt-Lundstadt, Smith, and Layton [4] found that the quality and quantity of individuals’ social relationships predicted morbidity and mortality above and beyond the Big 4. Specifically, the more socially disconnected people were, the more likely they were to develop CMDs, cancer, and mental health problems. The study accounted for age, gender, pre-existing health conditions, and SES. Lundstadt et al. [4] also argued that in spite of the evidence, weak social connectedness was grossly underestimated as a health-risk factor by health professionals and laypeople alike, and as such was seldomly addressed in healthcare settings. More recent studies indicate that little, if anything, has changed in this regard in the past decade [5, 6].

The discrepancy between knowledge and perception of weak social connectedness as a significant health-risk factor may reside in the notion that this construct–as it relates to physical health–is less intuitive in theory and practice compared to biological or behavioral risk factors. Haslam, Haslam, Jetten, Cruwys, and Bentley [7] also argue that the biopsychosocial model of health is overly reliant on biological health processes and tends to integrate psychological and social health determinants as secondary factors. Nonetheless, budding research has attempted to clarify the link between social connectedness and chronic somatic disease to improve the applicability of this knowledge in health care (e.g. in improving clinical risk assessment, CMD-preventive guidelines, and population-health initiatives) [5].

In theory, the mechanisms that underpin the relationship between social connectedness and physical health relate primarily to three distinct hypotheses: 1) social health norms that exist within social networks influence individual health behavior for better or for worse; 2) being embedded in a social network facilitates overall well-being, including physical health, through the availability of instrumental social support (e.g. financial support, work-life balance, health care); 3) being connected to others facilitates access to emotional and moral social support (“being there” for someone in times of need) which fosters individual psychological capital and resilience to everyday stress and the associated physiological wear and tear [810]. Thus, in terms of physical-health consequences, the former two hypotheses focus on the impact of social connectedness on CMD-risk factors (i.e. health norms, behavior, availability of concrete support), while the latter one centers more on biological endpoints (i.e. stress-related physiological reactivity that predicts CMDs). While these three mechanisms most certainly overlap to some extent (e.g. stress may lead to physiological reactivity as well as a propensity to engage in health-risk behaviors such as smoking), this review deals exclusively with the third pathway that focuses on the stress-buffering capacity of social connectedness and consequent physiological outcomes. We focus specifically on the toll of everyday stress (e.g. work-related stress, financial stress, social adversity) as opposed to acute, temporary stressors (e.g. the death of a loved one).

Strong scientific evidence shows a positive association between CMD and various types of everyday stress, including work-related stress [1113], stress related to financial adversity [14, 15], marital stress [15], caregiving [15], and the regular experience of discrimination [1620]. These associations remain strong even when controlling for demographic, social, and health-behavioral factors [21, 22, 23]. Emerging research has increasingly focused on allostatic load as the key physiological manifestation of stress that over time may lead to CMD [2429]. The term allostatic load (AL) was coined by McEwen and Stellar [30] who defined it as the cumulative physiological wear and tear that results from the regular experience of psychological stress. Specifically, hypothalamic-pituitary-adrenal (HPA) axis reactivity naturally fluctuates in response to external stressors–a phenomenon referred to as allostasis. In a healthy setting with no lingering stressors, the individual’s physiological system will oscillate naturally between allostasis and the baseline level of homeostasis as stress arises and abates. However, when stress becomes chronic, so too does allostasis persist over time, taxing the physiological system and ultimately resulting in AL [30]. Consistent with the evidence base on stress and CMD, the specific link between AL and serious illnesses–including CMDs and cancer–has been supported in a number of previous studies and reviews [8, 9, 13, 16, 17, 22, 3133], advancing the idea that AL is a central mediator of the physical health consequences of everyday psychological stress.

Combining the evidence on the relationship between social connectedness and physical health with the findings on stress, AL, and CMD, we argue that a key mechanism by which social connectedness protects individual health is by effectively buffering against AL. Indeed, since the early 2000s, studies explicitly testing this hypothesis have slowly, but steadily gained momentum and credibility. This research, however, spans decades and multiple methodologies and research designs, generating a somewhat disjointed empirical narrative. Thus, the aim of the present paper is to consolidate and synthesize the current evidence base on the association between social connectedness, stress, and AL.

Rationale

To the authors’ knowledge, few other reviews in this area have been conducted, and most (if not all) have focused on the broader relationship between for instance social support and cardiovascular markers of health (i.e. not necessarily stress-related reactivity). The gradual introduction of AL into the mainstream literature, however, has prompted a conceptual shift in the field by operationalizing stress as a psychophysiological phenomenon rather than a purely psychological one. Uchino et al. [9] conducted a highly informative review of the relationship between social support and physiological processes, parts of which focused on the impact of acute stress (not everyday stress) on physiological reactivity. Ten years on, Uchino [8] published a narrative review on this topic, again with sections on social support and physiological acute-stress reactions. More recently, Hostinar [34] and Wiley, Bei, Bower, and Stanton [35] also conducted reviews in this area. However, the former examines only literature published between 2013 and 2015, and the latter focuses on the relationship between AL and broad psychosocial measures rather than social connectedness per se. Thus, while we have identified four relevant reviews, two are 13 and 23 years old [8, 9], and none have explicitly focused on the link between social connectedness, everyday stress, and AL. In light of this, by critically assessing the entire body of research on this topic, the authors hope to make a significant addition to the field by updating and specifying past review efforts. Specifically, we will identify any relevant themes and variations that might have been missed in past reviews or appeared in subsequent research, qualitatively assess the evidence base, and highlight avenues for further study. Ultimately and most importantly, we believe that mapping out the link between social connectedness, stress, and CMD will help crystallize the health-protective mechanisms of social connectedness, and push it to the fore in clinical as well as research settings so it may be appropriately regarded as a crucial health-risk factor. Specifically, we argue that the integration of social connectedness measures in clinical health-assessment algorithms will improve algorithm accuracy and usability.

Method

Protocol

This review was conducted according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines. Full details can be accessed at www.prisma-guidelines.org.

Literature search strategy

A diverse range of scientific journals for literature on the relationship between social connectedness and AL was examined. Specifically, a comprehensive and rigorous search of the following EBSCOhost databases was executed: Academic Search Premier; AMED; Global Health; SocINDEX with Full Text; CINAHL Plus with Full Text; E-Journals; MEDLINE; Psychology and Behavioral Sciences Collection; SCOPUS; Science Direct. A literature search using Web of Knowledge and Google Scholar to identify any additional references that might have been missed was also performed. Additionally, reference lists of relevant papers were manually searched.

The predefined Boolean/phrase search terms related directly to the conceptualization of the key variables of interest: Social connectedness and allostatic load. For the sake of comprehensiveness, social connectedness was defined as an umbrella term, covering measures of number and/or quality of individual social ties, actual and perceived group memberships, and extent of social network. Measures of social support were also included as a proxy for social connectedness. In terms of AL, articles that dealt explicitly with this type of physiological reactivity represented the main focus. However, research that might not refer specifically to AL, but which nonetheless used AL biomarkers as stress-related outcomes was also included. For these latter studies, the Allostatic Load Index (ALI) was consulted to assess relevance. ALI is a commonly accepted measure of AL that comprises neuroendocrine and cardiovascular biomarkers, including systolic and diastolic blood pressure, total cholesterol, high-density-lipoprotein, glycosylated hemoglobin, waist-to-hip ratio, dehydroepiandrosterone sulfate, urinary epinephrine, norepinephrine, cortisol, as well as inflammatory markers such as C-reactive protein and cytokine levels [36].

The exact literature search syntax was as follows: “Allostasis” AND/OR “Allostatic load” AND/OR “Physiological reactivity” AND/OR “Psychophysiological reactivity” AND “Social connectedness” AND/OR “Group membership” AND/OR “Social identity” AND/OR “Social ties” AND/OR “Social network”. Search limiters were also specified to exclude obviously unrelated topics, such as for example papers on professional practice, philosophy, case studies, etc. Once the search had been executed, papers were selected based on the following inclusion criteria:

  1. The paper reported empirical studies on the relationship between human social connectedness, chronic and/or everyday (not acute) chronic stress, and AL.
  2. The full text was available.
  3. The paper was in English.
  4. The paper had undergone scientific peer review.
  5. The paper had been published since 1990.

Each database hit was evaluated by the authors in three rounds against the inclusion criteria. In the first round, papers that clearly did not relate to the subject matter were rejected (usually based on title). In the second round, abstracts of the papers retained in the first round were reviewed. Any article that failed to meet the inclusion criteria was discarded. Finally, the papers that remained after the first two evaluation rounds were downloaded and scrutinized in full-text detail for relevance. Only papers that passed through each of these three rounds were ultimately retained for inclusion in the review.

Research quality appraisal

The final round of search results evaluation included an appraisal of the retained papers’ methodological quality. To this end, the Quality Assessment Tool for Quantitative Studies (QATQS) was employed [37]. The QATQS evaluates research along six dimensions, including study population selection bias, study design, confounding variables, researcher blinding, data collection methods, and participant withdrawal and attrition. Each dimensional score for a given paper is averaged and combined into an overall assessment of the paper in terms of ‘weak’, ‘moderate’, or ‘strong’. The primary and secondary authors each coded all of the papers independently, and any differences in assessment were resolved through discussion and re-examination of the given paper.

Due to the small number of studies retained (n = 20), a meaningful relative weighting by study effect sizes (meta-analysis) was not feasible. Instead the strengths and weaknesses of each study were taken into account in a synthesis of overall findings in the discussion section.

Results

Literature search results

The initial search identified a total of 6022 articles. Of these, a subset of 144 papers were identified that in some way dealt with the convergence of social cognition/environment and physiological reactivity. Most of these articles were rejected due to one or more of the following reasons: The paper did not report on social connectedness as an independent, mediator, or moderator variable; the paper used composite measures of their social and physiological factors, preventing any isolated assessment of the variables of interest; the paper did not cite empirical research (e.g. editorial, comment); the paper reported insufficient statistical and/or methodological detail for assessment; the paper reported on animal populations; or a combination of these issues (see Fig 1). Ultimately, a total of 20 papers for were included in the present review. In the following sections, this evidence will be critically reviewed as it relates to the AL-buffering qualities of social connectedness.

Study characteristics and methodology

Of the 20 papers included in this review, 12 were from the 2010s, and the remaining eight articles were published in the 2000s. Further, the vast majority of articles was US-based (n = 17), with only one other nationality represented (Taiwan). The study populations varied somewhat in terms of ethnicity (African-Americans, Mexican-Americans, White Americans, Taiwanese) and age (teenagers to 100-year-olds). The average sample size across studies was N = 1024 (min = 54, max = 6729). In terms of methodology, 10 studies reported cross-sectional data, six were longitudinal studies, three were multi-method longitudinal/cross-sectional studies, and one was quasi-experimental. The average timeframe for longitudinal studies was about nine years (min = 2 years, max = 18 years, median = 10 years).

The social connectedness and AL measures employed varied somewhat across studies. Social connectedness was assessed by individual or composite measures of perceived social connectedness, social isolation, social support, social network size, group membership, social contact, social integration, and/or social ties. Beyond the sheer extent and nature of social connectedness, many studies also included measures of connection quality–for example in terms of relationship negativity or social network strain. When measuring established constructs such as social support or social network size and/or composition, 15 studies used validated scales (e.g. the Berkman-Syme Social Network Index), however five articles either did not report which scales had been employed, or constructed scales specifically for their studies. In terms of AL, 15 studies used either the ALI or a composite measure of AL that included many if not all of the ALI biomarkers. On the other hand, some studies focused on a select few AL indicators, such as C-reactive protein (CRP), fibrinogen, serum albumin levels [38, 39] (see Table 1).

Research methodology and quality

The inter-rater research quality assessments, conducted by the primary and secondary authors, were overall well-aligned. In the initial round of assessments, assessments deviated on only one qualitative dimension in each of three papers (81.25% inter-rater reliability). These discrepancies were resolved by reexamination of the given papers and subsequent discussion. Ultimately, scoring each paper along the six qualitative dimensions designated by the QATQS, 13 papers were deemed as based on ‘strong’ research, four as ‘moderate’, and three as ‘weak’ (see Table 1). The current evidence base on the relationship between social connectedness and AL thus rests on research of generally high methodological rigor.

Study findings

In the following paragraphs, a brief description of the main findings reported in the retained articles will be provided (see Table 1 for detailed information about study samples, design, and methodology).

In two studies, Brody et al. [40, 41] found that emotional support–derived from family, peer, and mentor networks–buffered against AL associated with neighborhood poverty and discrimination stressors. This effect remained even when controlling for a host of potentially confounding variables, including gender, health behavior, and SES. These findings were further backed up in similar research by Sotos-Prieto et al. [42] and Rosal et al. [43] who found that the extent of the individual’s social network and integration, and/or the quality of social support derived from their network, was inversely associated with AL. Indeed, the latter study found a 36% reduction in cortisol levels between those participants with the least amount social support versus those with the most.

A more extensive, ten-year longitudinal analysis of the association between social connectedness and AL, discovered a complex relationship that was moderated by several factors. These included age, the nature of ties with others (i.e. negative vs. positive), and the source of support received (spouse, friends, entire social network) [44]. Specifically, results indicated that negative and stressful relationships with a spouse and family members correlated positively and significantly with AL. Spousal support, on the other hand, correlated negatively with AL. Similarly, social network support correlated negatively with AL in young adults, while social network negativity correlated positively with AL. Counterintuitively, friend support and social network support were positively associated with AL in older adults, though the authors speculated that this might be due to greater feelings of social obligation in this subpopulation.

In another study, McClure et al. [45] investigated the effects of stress on Mexican immigrants to the US. Their findings indicated that the level of social support that women participants derived from their family networks was negatively correlated with AL. This relationship, however, was only apparent for women who lived in majority White communities as opposed to Mexican enclaves. Specifically, women who lived in White society and who had low family support were eight times more likely to experience everyday stress and associated AL than their high-support counterparts. These results thus complement those generated by Brooks et al. [44] in terms of connection quality and value of specifically familial social ties. However, McClure et al.’s findings also suggest that external factors such as the demography of one’s social environment may play a role in this relationship. These results should be interpreted with caution, though, as they are based on an exceedingly small sample size (N = 126).

Extending this research further, Seeman et al. [4648] made similar, but also more detailed discoveries that add more texture to the link between social connectedness and AL. In their initial, two-cohort study (Cohort 1: 70-79-year-olds, Cohort 2: 58-59-year-olds), the authors assessed the link between social connectedness, gender, and stress and AL. The results indicated several gender differences in the specific stress-related physiological reactivity that contributed to overall AL. In particular, across both cohorts, cardiovascular factors (e.g. BP, cholesterol, waist/hip ratio) were the most common contributors to AL in men. However, in women, neuroendocrine components (urinary cortisol, catecholamines) represented the main drivers of AL. In other words, everyday stress appeared to physically manifest in different ways as a function of gender [48]. Further, the results indicated a range of age- and gender-related differences in the extent to which social connectedness buffered against stress and the associated AL. Results for the younger cohort indicated that some ties were more significant for one gender than the other. For instance, social relationships based on intellectual/recreational connections appeared to protect women, but not men, against high AL. On the other hand, strong maternal ties were associated with lower AL for men, but not for women. Regardless of gender, however, the overall association in the younger cohort indicated that more and more positive social ties predicted lower AL scores. By contrast, in the older cohort, a higher number of social ties and increased frequency of emotional support decreased stress and AL only in men. Finally, negative spousal relationships (characterized by criticism and/or high demands) predicted higher AL in men, but not in women. Negative relationships with children, however, were associated with higher AL for both genders [48].

Comparable results were generated in two subsequent studies by Seeman et al. [46, 47]. A two-cohort study (Cohort 1: 54-70-year-olds, Cohort 2: >70-year-olds), conducted in Taiwan, found that the quantity of social connections available to the individual correlated negatively with AL across both cohorts [46]. They also found that for younger males (in Cohort 1), having a spouse predicted lower AL. Further, for both males and females in Cohort 1, negative relationships were associated with higher AL. This latter finding in particular was further supported in Seeman et al. [47] where they found that social strain (i.e. everyday stress associated with positive vs. negative ties) was the strongest, positive predictor of AL in a sample of 32-45-year-old men and women. Indeed, while initial analyses revealed that AL was negatively associated with both the number of social ties and the level of perceived available social support, both of these relationships were rendered non-significant when controlling for social strain. Taken together, these three studies by Seeman et al. [4648] complement those by Brooks et al. and McClure et al. by reiterating the notion that the content and quality of social connections is at least as important as their sheer number in terms of the buffer effect. These studies also align by highlighting gender and age differences in both physiological reactivity to everyday stress as well as the extent to which social connectedness buffers against such stress and AL.

The results generated by Seeman et al. [4648] resonate strongly with more recent research, which delves deeper into the link between social connectedness, stress/AL, and CMD [38, 39, 49]. In their first study, Yang et al. [39] operationalized AL in terms of C-reactive protein (CRP), fibrinogen, and albumin serum levels. These are all known markers of stress-related physiological inflammation and risk factors for various cancers and CVD. Reasoning that social disconnection leaves the individual more vulnerable to the effects of everyday stress, participants’ physiological data was correlated with self-report measures of the extent of social network integration (SNI, measuring size of social network and frequency of contact). Complementing past research, the results indicated an overall negative association between SNI and measures of stress and AL. These correlations, however, were more pronounced for men than for women. Specifically, adjusting for age and race, men who were socially isolated were 58% more likely than their socially connected counterparts to have elevated CRP levels and 94% more likely to have fibrinogen levels in the highest quartile. By contrast, socially isolated women were 38% more likely than socially connected women to record high-risk levels of fibrinogen. They also only had slightly and non-significantly elevated CRP levels. Mapping these results onto mortality, the results showed that everyday stress and the associated AL accounted for statistically significant proportions of the association between social isolation and all-cause mortality (14%), CVD mortality (12%), and cancer mortality (24%).

Next, in a similar study using a cross-sectional design, Yang et al. [38] again found significant and negative associations between SNI and stress/AL. With SNI values categorized from 1 = low to 4 = high, results showed that individuals who scored an SNI of 3 or lower, exhibited increasingly heavier stress and AL burdens than those who scored a 4. The pattern across SNI categories thus mimicked a dose-response relationship. In particular, participants with an SNI score of 3, 2, or 0–1 were 1.49 (95%CI 1.08, 2.06), 1.69 (95%CI 1.21, 2.36), and 2.35 (95%CI 1.62, 3.40) times more likely, respectively, to have elevated AL than those who scored a 4. Further, adjusting for education level, the SNI-AL relationship observed in the total sample, was significantly more pronounced for low-educated people. Within this subsample, compared to participants who had an SNI of 4, those who scored 2–3 were twice as likely to have elevated AL, while people who scored 0–1 were three times as likely. Finally, the association between SNI and AL was also stronger in non-White than White populations. While this association was statistically non-significant, the authors speculated that this was due lack of statistical power in their non-White sample (n = 157). This variation between SES levels may signify disproportionate access to other stress-buffering resources than social connectedness (e.g. high-SES individuals may have better financial security/rainy day funds to fall back on when needed).

In another two-wave (six-year) longitudinal study, Yang et al. [49] gauged AL by systolic blood pressure (SBP) and hypertension diagnoses. They used the same SNI measure as in their earlier work, but with an added social support component. Their results were again consistent with their previous research. There was an overall inverse correlation between SNI and both SBP (p < 0.01) and hypertension (p < 0.01). Further, individuals with the lowest level of SNI were nearly three times as likely as their high-SNI counterparts to be hypertensive (OR = 2.72, 95%CI 1.75, 4.21). The results from the longitudinal analyses showed that individuals with low social support at Wave 1 experienced a slight increase in SBP at Wave 2 (β = 0.03, p = 0.03). However, this effect became non-significant when SNI was added to the analysis. In terms of hypertension, respondents with the lowest SNI score had a 75.3% increase in risk from Wave 1 to Wave 2 (six years later). Results also suggested that this effect was partially mediated by social support, though this association was statistically non-significant.

Only a single quasi-experimental study was retained in the search results [50]. This study measured glucocorticoid sensitivity, cortisol levels, and cytokine production as AL indicators in two groups of parents: One group with a chronically ill child (n = 25) versus another with a healthy child (n = 25). The former parent group experienced significantly more stress than the latter. Social connectedness was assessed in terms of group membership and social support. There were no statistically significant between-group differences in cytokine production, however parents of chronically ill children showed significantly less dexamethasone suppression of IL-6 production than parents with healthy children. In terms of cortisol levels, parents with ill children had lower morning levels (p < .01) than their counterparts. Group membership and instrumental social support interacted to impact negatively and significantly on glucocorticoid sensitivity in parents with ill children, but in parents with healthy ones.

Finally, seven articles identified in the literature search reported weaker or null overall associations between the principal variables of interest [5157]. In a study focusing on the stress-buffering aspects of religious group memberships, Maselko et al. [52] (N = 853 70-80-year-olds) found no statistically significant relationship between the number of social ties and AL. They did, however, discover that weekly religious service attendance was associated with a 61% decreased risk of high AL for women, but not for men. The physiological reactivity was driven mainly by epinephrine levels and waist-hip ratio. In spite of their lacking results for social ties, the authors speculated that their measure of social connection and integration was too general in scope, and thus may have failed to capture any effects of social connectedness. Similarly, Weinstein et al. [53], Hawkley et al. [51], Gruenewald et al. [56], and Friedman et al. [57] investigated the relationship between SES-related stressors and AL. While each study found negative relationships between social connectedness measures an AL, none were statistically significant, with the exception of Weinstein who reported a significant positive association between widowhood (a marker of social connectedness) and AL (β = 0.24, p < .05). Lastly, Gersten [54] and Glei et al. [55] conducted separate studies in Taiwan. Both papers report negative, but weak relationships between social connectedness and AL.

Discussion

Main findings

Taken together, the results reported in the reviewed papers support the general conclusion that having ties to other persons or groups protects the individual against everyday stress and the associated AL. Controlling for a broad range of covariates, 13 out of 20 retained studies reported significant and inverse links between social connectedness, stress, and AL. The remaining seven studies consistently reported inverse, albeit non-significant associations between these factors [5157]. These results dovetail nicely with the extant literature on the strong positive association between social connectedness and physical health described in the introduction [46]. Specifically, the themes that emerge from the research further bolsters the notion that one way that social connectedness protects the individual from chronic disease is by buffering against stress and AL. Ultimately, this reiterates the fact that weak social connectedness is a major health-risk factor, and one that should be front and center in CMD- and cancer-preventive interventions, health-risk algorithms, clinical screening procedures, and the like.

While the overall relationship between social connectedness and AL is relatively clear, the pattern of underlying mechanisms that emerged in the review adds a layer of complexity. In particular, it was evident that gender, SES, as well as the quality and nature of the individual’s social connections often decide the extent to which social connectedness strengthens or weakens individual resilience to stress and AL. Specifically, and as one might expect, a negative and/or strained relationship seems to be more of a liability than an advantage [47], while positive and supportive relationships have the opposite effect [4648]. Furthermore, men appear to benefit more from spousal and parental ties as well as from general social connectedness than women do [39, 46, 48]. On the other hand, women seem to gain more than men from friendships and general familial relationships [45, 48]. Finally, SES markers also appeared to moderate the relationship, with the link between social connectedness and AL being stronger for low-educated and/or minority populations [38, 45]. While the reviewed evidence base is relatively small, the magnitude of the findings is augmented in the context of past research, indicating that the well-being effects of social connectedness are moderated by a variety of factors, including quality, content (e.g. social status and gender), and nature of individuals’ multiple social connections [5, 5860].

Strengths and limitations

In spite of its relatively small size in terms of sheer study quantity, the reviewed evidence base has a number of redeeming strengths. As indicated in the results section, the systematic quality assessment of the literature suggests that the studies discussed here are mostly of high empirical quality. Only three of 20 studies were deemed as methodologically “weak”, with most of the rest coming out as “strong”. These high ratings are mainly due to the rigorous research designs employed, the mostly large and representative study populations, as well as the effort to control for a broad range of potentially extraneous variables (see Table 1). In particular, the longitudinal designs employed by Brody et al., Brooks et al., Seeman et al., Weinstein et al., and Yang et al. are noteworthy as strong approaches to the measurement of the link between social connectedness and AL. Most of these studies tapped relatively large and representative populations, often across a decade or more, and as a result generated highly compelling findings and conclusions.

Another strength relates to the physiological nature of the main outcome variable. Most research on the costs and benefits of social connectedness rely on self-report scales of well-being and health, and as such is prone to the considerable bias and error associated with subjective measures of this sort. As mentioned in the introduction, this may be part of the reason why weak social connectedness, in spite of the evidence, is treated by clinicians and researchers as a second-tier health-risk factor. The reviewed studies, however, circumvent many of these validity issues of self-report measures by focusing on purely physiological outcomes of stress and their relation to social connectedness. This expounds the very real transference of negative psychological experience to physical health-risk factors.

The use of broad, multi-component AL measures in most of the studies also speaks to the level of methodological rigor employed. The pertinence of measuring multiple markers of AL is especially evident when considering the gender differences in the physiological manifestation of stress seen in Yang et al. [39]. In other words, inclusive AL measures may be necessary to capture any variation in reactivity that may exist across different populations.

There are also a few limitations that need to be addressed. First, all of the studies included in the present review were correlational in nature, and thus cannot account for any causal relationships. In the context of relevant past experimental research, however, the causal impact of social connectedness on physiological stress reactivity may be surmised with some level of probability. For instance, several laboratory paradigms have demonstrated the protective qualities of social connectedness by experimentally manipulating social support during acute stress. Typically, these studies indicated that the physical or mental presence of social support (e.g. a friend or confederate accompanying the participant; making social connectedness mentally salient) during acute stress minimized the physiological reactivity to that stress compared to no/minimal-support control groups [8, 9, 6064].

Another limitation relates to the fact that all but three studies were based in the US, and only four studies focused on non-Western populations (Taiwanese). This is a considerable shortcoming as past research has found cultural differences (e.g. between Western individualist and Eastern collectivist culture) in resilience to stressors as well as in the nature, structure, and perceived value of social networks and relationships [65]. Given these differences, it is conceivable that the relationship between social connectedness and AL observed in Western studies may play out differently in non-Western populations. Indeed, the Taiwanese studies were disproportionately more likely to report null associations between social connectedness and AL. Specifically, of a total four studies on Taiwanese populations, three indicated inconclusive or non-significant results. Thus, cross-cultural comparisons could potentially provide deeper insight into the social connectedness-AL relationship.

In addition, while the frequent use of multi-component AL measures represents an overall strength in the reviewed research, the cross-study variability in AL indicators may nonetheless complicate efforts to synthesize and compare the results of the different papers. For example, while some studies focused on neuroendocrine markers of AL, others focused on cardiometabolic indices. Future research might mitigate this issue by planning studies around the Allostatic Load Index (ALI) [36] where possible.

Similarly, the variation across studies in the measurement of social connectedness also represents a potential limitation to the literature. Most of the reviewed studies operationalized connectedness by number and characteristics of individual ties and relationships. However, other research indicates that conceptualizing social connection in terms of group memberships rather than interpersonal ties is a stronger predictor of health and well-being [66]. Specifically, this research indicates that the groups with which we identify (e.g. gender, ethnicity, profession, hobby, etc.) provide us with a variety of shared social identities that enable the individual to conceptualize the self in terms of their group belonging (“us” and “we”) rather than merely a personal identity (“me” and “I”). Faced with adversity, this sense of group-based community and solidarity may facilitate individual resilience and psychological capital [5, 7, 60, 6668]. Additionally, because social identities are domain specific they allow the individual to shift away from a stressful identity (e.g. one’s job) and tap into other, more positive and harmonious ones for respite and support (e.g. hobby, religion) [5, 59, 60, 69]. Given these findings, future research should focus on whether multiple group membership buffers against AL in a similar fashion as interpersonal relationships do.

The issue of our broad definition of social connectedness, follows naturally from the previous point. As noted in the methods section, we deliberately accepted an operationalization of social connectedness that included various measures of, for instance, social network size, quality, and content, social support, social activity and integration, etc. Thus, the reviewed studies are somewhat lacking in terms of a clear definition of social connectedness. Nonetheless, consistent with past research, our results strongly indicate that it is both the extent and quality of social connectedness that protect the individual against AL. Hence, in terms of applying social connectedness as a screenable health-risk factor, this suggests that in addition to gauging the mere extent or presence of an individual’s social network, clinical measures should also focus on the quality (e.g. social support, integration, activity) of that network.

Finally, as mentioned previously, the evidence base is quite small. For this reason, only the overall negative relationship between social connectedness and AL stood out with any clarity. However, this result may be skewed by publication bias favoring significant results. Further, the present review identified several potential mechanisms that may underpin the association between social connectedness and AL. These relate mainly to gender, social status, and quality and nature of social connections. However, more research drilling down into the true significance of these particular factors is needed.

Conclusion and future directions

The present review advances the notion of weak social connectedness as a significant health-risk factor. It does so by emphasizing the psychophysiological mechanism by which social connectedness moderates everyday stress and thus AL. The bigger and more supportive the individual’s social network is, the lower the likelihood that she/he will experience AL. Understanding the underlying mechanisms of this relationship opens up promising avenues for primary-care interventions targeting CMD-risk factors. It also stresses that weak social connectedness could and should be regarded in clinical settings as a significant health-risk factor. While this review highlights the negative overall association between social connectedness and AL, the current evidence base is limited in size and scope. Firm conclusions about the veracity of moderators identified in individual studies therefore remain elusive. In the context of applying measures of social connectedness in preventive clinical settings, these shortcomings in the evidence warrant further research to totally clarify when and how social connectedness impacts on AL. Specifically, the present review highlights gender, SES, and cultural variation as potential, but under-researched moderators. The need to unravel the numerous ways of conceptualizing and measuring social connectedness are also emphasized. The authors hope that this review will serve as a jumping-off point for future research to unpack the total relationship between social connectedness, AL, and CMD.

Supporting information

References

  1. 1. WHO. Noncommunicable diseases country profiles 2018. Geneva; 2018.
  2. 2. Loef M, Walach H. The combined effects of healthy lifestyle behaviors on all cause mortality: a systematic review and meta-analysis. Preventive medicine. 2012;55(3):163–70. pmid:22735042
  3. 3. Holt-Lunstad J. The potential public health relevance of social isolation and loneliness: Prevalence, epidemiology, and risk factors. Public Policy & Aging Report. 2017;27(4):127–30.
  4. 4. Holt-Lunstad J, Smith TB, Layton JB. Social relationships and mortality risk: a meta-analytic review. PLoS medicine. 2010;7(7):e1000316. pmid:20668659
  5. 5. Haslam C, Jetten J, Cruwys T, Dingle G, Haslam SA. The New Psychology of Health. Unlocking the Social Cure. 2018:2018.
  6. 6. Leigh-Hunt N, Bagguley D, Bash K, Turner V, Turnbull S, Valtorta N, et al. An overview of systematic reviews on the public health consequences of social isolation and loneliness. Public Health. 2017;152:157–71. pmid:28915435
  7. 7. Haslam SA, Haslam C, Jetten J, Cruwys T, Bentley S. Group life shapes the psychology and biology of health: The case for a sociopsychobio model. Social and Personality Psychology Compass. 2019:e12490.
  8. 8. Uchino BN. Social support and health: a review of physiological processes potentially underlying links to disease outcomes. Journal of behavioral medicine. 2006;29(4):377–87. pmid:16758315
  9. 9. Uchino BN, Cacioppo JT, Kiecolt-Glaser JK. The relationship between social support and physiological processes: a review with emphasis on underlying mechanisms and implications for health. Psychological bulletin. 1996;119(3):488. pmid:8668748
  10. 10. Uchino BN, Smith TW, Carlisle M, Birmingham WC, Light KC. The Quality of Spouses’ Social Networks Contributes to Each Other’s Cardiovascular Risk. PLoS ONE. 2013;8(8):1–7.
  11. 11. Kivimäki M, Jokela M, Nyberg ST, Singh-Manoux A, Fransson EI, Alfredsson L, et al. Long working hours and risk of coronary heart disease and stroke: a systematic review and meta-analysis of published and unpublished data for 603,838 individuals. Lancet. 2015;386 North American Edition(10005):1739–46. pmid:26298822
  12. 12. Kivimäki M, Nyberg ST, Batty GD, Fransson EI, Heikkilä K, Alfredsson L, et al. Job strain as a risk factor for coronary heart disease: a collaborative meta-analysis of individual participant data. The Lancet. 2012;380(9852):1491–7.
  13. 13. Kivimäki M, Steptoe A. Effects of stress on the development and progression of cardiovascular disease. Nature Reviews Cardiology. 2018;15(4):215. pmid:29213140
  14. 14. Evans GW, Kim P. Childhood poverty and health: cumulative risk exposure and stress dysregulation. Psychological Science. 2007;18(11):953–7. pmid:17958708
  15. 15. Dimsdale JE. Psychological stress and cardiovascular disease. Journal of the American College of Cardiology. 2008;51(13):1237–46. pmid:18371552
  16. 16. Geronimus AT, Hicken M, Keene D, Bound J. “Weathering” and age patterns of allostatic load scores among blacks and whites in the United States. American journal of public health. 2006;96(5):826–33. pmid:16380565
  17. 17. Troxel WM, Matthews KA, Bromberger JT, Sutton-Tyrrell K. Chronic stress burden, discrimination, and subclinical carotid artery disease in African American and Caucasian women. Health Psychology. 2003;22(3):300. pmid:12790258
  18. 18. Chae DH, Lincoln KD, Adler NE, Syme SL. Do experiences of racial discrimination predict cardiovascular disease among African American men? The moderating role of internalized negative racial group attitudes. Social Science & Medicine. 2010;71(6):1182–8.
  19. 19. Forsyth J, Schoenthaler A, Chaplin WF, Ogedegbe G, Ravenell J. Perceived discrimination and medication adherence in black hypertensive patients: the role of stress and depression. Psychosomatic Medicine. 2014;76(3):229–36. pmid:24677163
  20. 20. Brondolo E, Love EE, Pencille M, Schoenthaler A, Ogedegbe G. Racism and hypertension: a review of the empirical evidence and implications for clinical practice. American journal of hypertension. 2011;24(5):518–29. pmid:21331054
  21. 21. Batten SV, Aslan M, Maciejewski PK, Mazure CM. Childhood maltreatment as a risk factor for adult cardiovascular disease and depression. Journal of Clinical Psychiatry. 2004;65(2):249–54. pmid:15003081
  22. 22. Danese A, McEwen BS. Adverse childhood experiences, allostasis, allostatic load, and age-related disease. Physiology & behavior. 2012;106(1):29–39.
  23. 23. Scott KM, Von Korff M, Angermeyer MC, Benjet C, Bruffaerts R, De Girolamo G, et al. Association of childhood adversities and early-onset mental disorders with adult-onset chronic physical conditions. Archives of general psychiatry. 2011;68(8):838–44. pmid:21810647
  24. 24. Logan JG, Barksdale DJ. Allostasis and allostatic load: expanding the discourse on stress and cardiovascular disease. Journal of clinical nursing. 2008;17(7b):201–8. pmid:18578796
  25. 25. Low CA, Salomon K, Matthews KA. Chronic life stress, cardiovascular reactivity, and subclinical cardiovascular disease in adolescents. Psychosomatic medicine. 2009;71(9):927. pmid:19737856
  26. 26. McEwen BS. Allostasis and allostatic load: implications for neuropsychopharmacology. Neuropsychopharmacology. 2000;22(2):108–24. pmid:10649824
  27. 27. Seeman TE, McEwen BS, Rowe JW, Singer BH. Allostatic load as a marker of cumulative biological risk: MacArthur studies of successful aging. Proceedings of the National Academy of Sciences. 2001;98(8):4770–5.
  28. 28. Tiedt JA, Brown LA. Allostatic load: The relationship between chronic stress and diabetes in Native Americans. Journal of Theory Construction & Testing. 2014;18(1):22.
  29. 29. Vines AI, Ward JB, Cordoba E, Black KZ. Perceived racial/ethnic discrimination and mental health: A review and future directions for social epidemiology. Current epidemiology reports. 2017;4(2):156–65. pmid:28920011
  30. 30. McEwen BS, Stellar E. Stress and the individual: mechanisms leading to disease. Archives of internal medicine. 1993;153(18):2093–101. pmid:8379800
  31. 31. Beckie TM. A systematic review of allostatic load, health, and health disparities. Biological research for nursing. 2012;14(4):311–46. pmid:23007870
  32. 32. Steptoe A, Kivimäki M. Stress and cardiovascular disease: an update on current knowledge. Annual review of public health. 2013;34:337–54. pmid:23297662
  33. 33. Steptoe A, Kivimäki M. Stress and cardiovascular disease. Nature Reviews Cardiology. 2012;9(6):360. pmid:22473079
  34. 34. Hostinar CE. Recent developments in the study of social relationships, stress responses, and physical health. Current opinion in psychology. 2015;5:90–5. pmid:26366429
  35. 35. Wiley JF, Bei B, Bower JE, Stanton AL. Relationship of psychosocial resources with allostatic load: a systematic review. Psychosomatic medicine. 2017;79(3):283. pmid:27768647
  36. 36. Mauss D, Li J, Schmidt B, Angerer P, Jarczok MN. Measuring allostatic load in the workforce—a systematic review. Industrial Health. 2014.
  37. 37. Thomas B, Ciliska D, Dobbins M, Micucci S. A process for systematically reviewing the literature: providing the research evidence for public health nursing interventions. Worldviews on Evidence‐Based Nursing. 2004;1(3):176–84. pmid:17163895
  38. 38. Yang YC, Li T, Frenk SM. Social network ties and inflammation in US adults with cancer. Biodemography and social biology. 2014;60(1):21–37. pmid:24784985
  39. 39. Yang YC, McClintock MK, Kozloski M, Li T. Social isolation and adult mortality: the role of chronic inflammation and sex differences. Journal of health and social behavior. 2013;54(2):183–203. pmid:23653312
  40. 40. Brody GH, Lei M-K, Chen E, Miller GE. Neighborhood poverty and allostatic load in African American youth. Pediatrics. 2014;134(5):e1362–e8. pmid:25311604
  41. 41. Brody GH, Lei MK, Chae DH, Yu T, Kogan SM, Beach SR. Perceived Discrimination Among A frican A merican Adolescents and Allostatic Load: A Longitudinal Analysis With Buffering Effects. Child development. 2014;85(3):989–1002. pmid:24673162
  42. 42. Sotos-Prieto M, Bhupathiraju SN, Falcón LM, Gao X, Tucker KL, Mattei J. A healthy lifestyle score is associated with cardiometabolic and neuroendocrine risk factors among Puerto Rican adults. The Journal of nutrition. 2015;145(7):1531–40. pmid:25948783
  43. 43. Rosal MC, King J, Ma Y, Reed GW. Stress, social support, and cortisol: inverse associations? Behavioral Medicine. 2004;30(1):11–22. pmid:15473629
  44. 44. Brooks KP, Gruenewald T, Karlamangla A, Hu P, Koretz B, Seeman TE. Social relationships and allostatic load in the MIDUS study. Health Psychology. 2014;33(11):1373. pmid:24447186
  45. 45. McClure HH, Snodgrass JJ, Martinez CR, Squires EC, Jiménez RA, Isiordia LE, et al. Stress, place, and allostatic load among Mexican immigrant farmworkers in Oregon. Journal of immigrant and minority health. 2015;17(5):1518–25. pmid:25724150
  46. 46. Seeman T, Glei D, Goldman N, Weinstein M, Singer B, Lin Y-H. Social relationships and allostatic load in Taiwanese elderly and near elderly. Social science & medicine. 2004;59(11):2245–57.
  47. 47. Seeman TE, Gruenewald TL, Cohen S, Williams DR, Matthews KA. Social relationships and their biological correlates: Coronary Artery Risk Development in Young Adults (CARDIA) study. Psychoneuroendocrinology. 2014;43:126–38. pmid:24703178
  48. 48. Seeman TE, Singer BH, Ryff CD, Love GD, Levy-Storms L. Social relationships, gender, and allostatic load across two age cohorts. Psychosomatic medicine. 2002;64(3):395–406. pmid:12021414
  49. 49. Yang YC, Boen C, Mullan Harris K. Social relationships and hypertension in late life: evidence from a nationally representative longitudinal study of older adults. Journal of aging and health. 2015;27(3):403–31. pmid:25253728
  50. 50. Miller GE, Cohen S, Ritchey AK. Chronic psychological stress and the regulation of pro-inflammatory cytokines: a glucocorticoid-resistance model. Health psychology. 2002;21(6):531. pmid:12433005
  51. 51. Hawkley LC, Lavelle LA, Berntson GG, Cacioppo JT. Mediators of the relationship between socioeconomic status and allostatic load in the Chicago Health, Aging, and Social Relations Study (CHASRS). Psychophysiology. 2011;48(8):1134–45. pmid:21342206
  52. 52. Maselko J, Kubzansky L, Kawachi I, Seeman T, Berkman L. Religious service attendance and allostatic load among high-functioning elderly. Psychosomatic Medicine. 2007;69(5):464–72. pmid:17567709
  53. 53. Weinstein M, Goldman N, Hedley A, Yu-Hsuan L, Seeman T. Social linkages to biological markers of health among the elderly. Journal of Biosocial Science. 2003;35(3):433–53. pmid:12887222
  54. 54. Gersten O. Neuroendocrine biomarkers, social relations, and the cumulative costs of stress in Taiwan. Social Science & Medicine. 2008;66(3):507–19.
  55. 55. Glei DA, Goldman N, Chuang Y-L, Weinstein M. Do chronic stressors lead to physiological dysregulation? Testing the theory of allostatic load. Psychosomatic Medicine. 2007;69(8):769–76. pmid:17942833
  56. 56. Gruenewald TL, Karlamangla AS, Hu P, Stein-Merkin S, Crandall C, Koretz B, et al. History of socioeconomic disadvantage and allostatic load in later life. Social science & medicine. 2012;74(1):75–83.
  57. 57. Friedman EM, Karlamangla AS, Gruenewald T, Koretz B, Seeman TE. Early life adversity and adult biological risk profiles. Psychosomatic medicine. 2015;77(2):176. pmid:25650548
  58. 58. Brook AT, Garcia J, Fleming MA. The effects of multiple identities on psychological well-being. Personality and Social Psychology Bulletin. 2008;34(12):1588–600. pmid:19050334
  59. 59. Iyer A, Jetten J, Tsivrikos D, Postmes T, Haslam SA. The more (and the more compatible) the merrier: Multiple group memberships and identity compatibility as predictors of adjustment after life transitions. British Journal of Social Psychology. 2009;48(4):707–33.
  60. 60. Sønderlund AL, Morton TA, Ryan MK. Multiple group membership and well-being: Is there always strength in numbers? Frontiers in psychology. 2017;8:1038. pmid:28680414
  61. 61. Gerin W, Pieper C, Levy R, Pickering TG. Social support in social interaction: a moderator of cardiovascular reactivity. Psychosomatic medicine. 1992;54(3):324–36. pmid:1620808
  62. 62. Grant N, Hamer M, Steptoe A. Social isolation and stress-related cardiovascular, lipid, and cortisol responses. Annals of Behavioral Medicine. 2009;37(1):29–37. pmid:19194770
  63. 63. Lepore SJ. Cynicism, social support, and cardiovascular reactivity. Health Psychology. 1995;14(3):210. pmid:7641661
  64. 64. Lepore SJ, Allen K, Evans GW. Social support lowers cardiovascular reactivity to an acute stressor. Psychosomatic Medicine. 1993;55(6):518–24. pmid:8310112
  65. 65. Townsend KC, McWhirter BT. Connectedness: A review of the literature with implications for counseling, assessment, and research. Journal of Counseling & Development. 2005;83(2):191–201.
  66. 66. Jetten J, Branscombe NR, Haslam SA, Haslam C, Cruwys T, Jones JM, et al. Having a lot of a good thing: Multiple important group memberships as a source of self-esteem. PloS one. 2015;10(5):e0124609. pmid:26017554
  67. 67. Haslam C, Jetten J, Alexander SH. The Social Cure: Identity, Health and Well-being: Psychology Press; 2012.
  68. 68. Jetten J, Haslam C, Alexander SH. The social cure: Identity, health and well-being: Psychology Press; 2012.
  69. 69. Jones JM, Jetten J. Recovering from strain and enduring pain: Multiple group memberships promote resilience in the face of physical challenges. Social Psychological and Personality Science. 2011;2(3):239–44.