Skip to main content
PMC home page
The Journals of Gerontology Series B: Psychological Sciences and Social Sciences logoLink to The Journals of Gerontology Series B: Psychological Sciences and Social Sciences
. 2019 Sep 9;75(10):2132–2141. doi: 10.1093/geronb/gbz116

The Prospective Association of Social Integration With Life Span and Exceptional Longevity in Women

Claudia Trudel-Fitzgerald 1,2,✉, Emily S Zevon 1, Ichiro Kawachi 1, Reginald D Tucker-Seeley 3, Francine Grodstein 4,5, Laura D Kubzansky 1,2
Editor: Lynn Martire
PMCID: PMC7664314  PMID: 31495897

Abstract

Objectives

Although stronger social relationships have been associated with reduced mortality risk in prior research, their associations with favorable health outcomes are understudied. We evaluated whether higher social integration levels were associated with longer life span and greater likelihood of achieving exceptional longevity.

Method

Women from the Nurses’ Health Study completed the Berkman–Syme Social Network Index in 1992 (N = 72,322; average age = 58.80 years), and were followed through 2014 with biennial questionnaires. Deaths were ascertained from participants’ families, postal authorities, and death registries. Accelerated failure time models adjusting for relevant covariates estimated percent changes in life span associated with social integration levels; logistic regressions evaluated likelihood of surviving to age 85 years or older among women who could reach that age during follow-up (N = 16,818).

Results

After controlling for baseline demographics and chronic diseases, socially integrated versus isolated women had 10% (95% confidence interval [CI] = 8.80–11.42) longer life span and 41% (95% CI = 1.28–1.54) higher odds of surviving to age 85 years. All findings remained statistically significant after further adjusting for health behaviors and depression.

Discussion

Better social integration is related to longer life span and greater likelihood of achieving exceptional longevity among midlife women. Findings suggest social integration may be an important psychosocial asset to evaluate for promoting longer, healthier lives.

Keywords: Death, Health, Mortality, Relationships, Social isolation

As life span has increased in industrialized countries, exceptional longevity—typically defined as survival to 85 years (Newman & Murabito, 2013; Revelas et al., 2018)—has become increasingly common. Empirical evidence obtained across diverse organisms has consistently demonstrated that improvements in life span often co-occur with delayed morbidity (Longo et al., 2015), indicating that studying factors associated with increased longevity may yield new insights regarding how to promote both long and healthy lives (also known as “healthspan”) (López-Otín, Blasco, Partridge, Serrano, & Kroemer, 2013). Research on exceptional longevity has largely focused on identifying biomedical factors (e.g., genetic variants) that are associated with increased survival, but an emerging body of research has suggested nongenetic factors matter as well. Accordingly, research has begun to identify psychosocial assets, such as optimism and other facets of psychological well-being, as potential predictors of longer life (Kubzansky et al., 2018; Lee et al., 2019; Steptoe, 2019).

Social relationships have also been identified as a key predictor of human health (Berkman & Krishna, 2014; Holt-Lunstad, Robles, & Sbarra, 2017). Research has demonstrated beneficial effects of social support and networks on a wide range of health outcomes (Berkman & Krishna, 2014; Holt-Lunstad et al., 2017; Trudel-Fitzgerald, Chen, Singh, Okereke, & Kubzansky, 2016), with cognitive, emotional, behavioral, and direct biological pathways proposed to explain observed associations (Berkman, Glass, Brissette, & Seeman, 2000; Cohen, 1988; Kroenke, 2018). The relationship between social relationships and premature mortality has been assessed extensively, with many studies demonstrating an association between social isolation and increased risk of premature death (Berkman & Krishna, 2014; Holt-Lunstad et al., 2017). This work has generally followed a traditional adverse-outcomes-oriented and risk-focused perspective. However, investigators have called for applying a positive health framework to gain greater insight into how to promote and preserve healthy functioning (Lloyd-Jones, 2014; National Research Council Committee on Future Directions for Behavioral and Social Sciences Research at the National Institute of Health, 2001). This shift in priorities grows out of an increasing understanding that insights derived from considering risk factors associated with increased disease and mortality may differ from those derived from examining positive factors that may be associated with the attainment and maintenance of good health. Moreover, the absence of a harmful risk factor is not necessarily the opposite of the presence of a positive or protective factor. For example, not being socially isolated is different from being social integrated. Depending on the measure used to assess social isolation, it may not be possible to determine whether an individual who does not meet criteria for social isolation is in fact truly socially integrated, without additional information. Recent studies investigating the potential role of positive factors with future risk of chronic diseases and premature mortality have found meaningful associations independent of not only conventional risk factors (e.g., health status), but also psychosocial risk factors (e.g., depressive symptoms), reinforcing the idea that positive factors capture more than merely the absence of negative factors (VanderWeele et al., in press). Yet, to the best of our knowledge no studies have explicitly examined the association of social relationships with exceptional longevity.

Research suggests being social integrated (and other psychosocial assets) is associated with health outcomes above and beyond the effects of other risk factors (Berkman & Krishna, 2014; Holt-Lunstad et al., 2017; Steptoe, 2019). From a positive health framework, social relationships are considered as a health asset or a life skill (Steptoe & Wardle, 2017), not only reducing likelihood of specific diseases, but also leading to positive health outcomes, such as achieving or maintaining health or healthy aging, more likely. Identifying diverse assets that promote health across the life course, particularly health in aging, will help inform efforts not only to reduce exposure to health risks, but also to achieve optimal functioning, informing a “primordial prevention” approach (Strasser, 1978). Although healthy aging is a multidimensional construct that is often defined to incorporate physical, cognitive, and emotional well-being, the achievement of long life span is its most basic prerequisite (Anton et al., 2015; Woods et al., 2016). By understanding assets that promote longevity, we can take a step outside the paradigm of disease and death, and create new insights regarding the means through which long and healthy lives can be achieved.

In this study, we assessed social integration, which refers to the number, type, and frequency of social contacts, and evaluated its association with increased longevity. We focus on social integration because it has been associated with health outcomes more consistently than other social relationship constructs such as emotional social support (Cohen & Janicki-Deverts, 2009; Holt-Lunstad et al., 2017; Nausheen, Gidron, Peveler, & Moss-Morris, 2009). We used data from the Nurses’ Health Study (NHS), a large ongoing cohort of women, to evaluate if higher levels of social integration are associated with longer life span, as well as with greater likelihood of attaining exceptional longevity. All analyses controlled for potential confounders, including demographics and initial health status, following prior research in this domain.

We also considered the role of depression, because prior research has suggested that greater social integration is associated with less depression (Chang, Pan, Kawachi, & Okereke, 2016) and also that depression is related to a greater risk of premature mortality (Wei et al., 2019). The direction of causality between social integration and depression is uncertain (Berkman et al., 2000), but due to data limitations, we considered depression as a potential confounder in sensitivity analyses. As lifestyle factors are posited to lie on the pathway linking social integration to longevity, we included a separate set of models adjusting for health-related behaviors to examine explicitly whether adding these variables may partly or fully explain the associations of interest. Finally, in secondary analyses we investigated individual domains of social integration (e.g., religious participation; number of close friends/relatives) to ascertain whether some components were differentially salient for longevity. We hypothesized that greater levels of social integration were associated with longer life span, as well as with greater likelihood of attaining exceptional longevity, beyond statistical adjustment for potential confounders and pathways.

Method

Study Population

Data are from the ongoing NHS cohort, which began in 1976 with 121,700 married female registered nurses aged 30–55 years old. Since 1976, NHS participants have returned biennial questionnaires collecting data on health, nutrition, and lifestyle, as well as a variety of social and psychological factors, with follow-up rate approximately 85%–90% (Bao et al., 2016). The sample for our primary analysis included women who completed the 1992 measure of social integration—the Berkman–Syme Social Network Index (SNI)—and have been followed through 2014. Participants were excluded from analyses if they were missing data on social integration or demographic covariates (excluding husband’s education for which we created a missing indicator because 15.51% of data were missing) or if they died within 2 years after baseline, to reduce likelihood of reverse causation whereby imminent death would influence social relationships or the reporting of them. These exclusions reduced the sample size from 103,601 to 72,322 women. This sample size is either comparable to or larger than most prior studies investigating either social integration with various health-related outcomes (Pinquart & Duberstein, 2010; Trudel-Fitzgerald et al., 2016) or psychosocial factors with longevity (Costa, Weiss, Duberstein, Friedman, & Siegler, 2014; Lee et al., 2019). Further assessment of the statistical stability of the results for this study was quantitatively evaluated by considering the width of confidence intervals (CIs). For analyses assessing the likelihood of survival to the age of 85 years, the sample was further restricted to participants born before 1928, for whom it was possible to reach the age of 85 years during the study period (N = 16,818). The study protocol was approved by the institutional review boards of the Brigham and Women’s Hospital.

Measures

Social integration

Social integration, a construct that captures the number, type, and frequency of social contacts, was assessed with the Berkman–Syme SNI (Berkman & Krishna, 2014; Berkman & Syme, 1979), administered via self-reported scale in 1992. The SNI assesses quantity and type of social relationships across four domains: marriage, contacts with close friends and relatives, participation in religious activities, and participation in group associations (Berkman & Syme, 1979). The measure has shown good test–retest reliability and acceptable construct validity, and has predicted breast cancer survival and mental functioning in NHS women (Trudel-Fitzgerald et al., 2016). Following prior work using the SNI in this cohort (Chang et al., 2017; Kroenke, Kubzansky, Schernhammer, Holmes, & Kawachi, 2006; Trudel-Fitzgerald et al., 2016), each of the four domains of social integration was scored from 0 (least integrated) to 3 (most integrated; Supplementary Table 1). These domain scores were summed to create a continuous SNI score ranging from 0 (highly socially isolated) to 12 (highly socially integrated). The continuous SNI score was considered missing if scores for any of the four domains were unavailable. This score was then divided into quartiles to allow for examination of potential discontinuous or threshold effects. Thus, participants were classified according to four levels of social integration: highly socially isolated (reference group), moderately isolated, moderately integrated, and highly socially integrated (Chang et al., 2017; Kroenke et al., 2006; Trudel-Fitzgerald et al., 2016). In the NHS, the SNI was administered every 4 years, covering a 16-year period from 1992 to 2008. Scores are fairly stable across assessments, as supported by a high intra-class correlation coefficient (ICC) value (ICC: 0.76, 95% CI: 0.76–0.77) and low within-subject variability (0.18, 95% CI: 0.18–0.18) in the current analytic sample. Therefore, we did not conduct additional analyses updating the SNI score or considering trajectories of change in SNI over time. Of note, to enter the cohort study when it was first initiated in 1976, women had to be married; consequently, most participants were still married or in a domestic partnership in 1992 when the SNI was first queried.

Life span

Life span was operationalized as changes in predicted life span. We also considered exceptional longevity, which was defined as survival to the age of 85 years or older. Deaths are reported by participants’ families and by postal authorities. The names of nonrespondents are searched within the National Death Index, which has compiled data from state death registries since 1979 and correctly identifies 98% of known deaths among a sample of NHS participants for whom death certificates were available (Rich-Edwards, Corsano, & Stampfer, 1994). Date of death is ascertained from death records. In this study, deaths were identified through the end of 2014, the most recently available data.

Covariates

All covariates were queried at baseline (in 1992), unless otherwise noted. Demographic variables including age (continuous), education level (registered nurse vs undergraduate/graduate degree), and husband’s education level (≤high school, above high school, missing level) were considered as potential confounders. Analyses also considered self-reported prevalence or history of the following major chronic diseases, individually (yes vs no): high cholesterol, high blood pressure, diabetes, cancer, stroke, and myocardial infarction (MI). Depressive symptoms were assessed via the five-item Mental Health Inventory (MHI-5), a subscale of the 36-Item Short Form Survey from the RAND Medical Outcomes Study (Ware & Sherbourne, 1992). Scores ranged from 0 (most depressed) to 100 (least depressed), and participants were classified as having clinical depressive symptoms (yes vs no) if their score was less than or equal to 60 (Rumpf, Meyer, Hapke, & John, 2001).

Health behavior-related variables, such as smoking status, physical activity, alcohol consumption, diet quality, and body mass index (BMI) were considered as covariates that might either confound or potentially mediate the association between social integration and longevity. Self-reported smoking status was defined as never, former, or current smoker. Physical activity was modeled as a dichotomous variable indicating whether the participant met recommended levels of physical activity (i.e., reporting ≥150 min of moderate-to-vigorous physical activity per week; yes vs no). Alcohol consumption and diet quality were assessed via a food frequency questionnaire (Yuan et al., 2017) administered in 1994. Alcohol was modeled as a dichotomous variable indicating whether participants met recommendations for no more than one serving of an alcoholic drink per day (yes vs no) (U.S. Department of Health and Human Services & U.S. Department of Agriculture, 2010). Diet quality was operationalized as a continuous variable using the Alternative Health Eating Index (AHEI), which assigns a dietary score ranging from 0 (lowest quality) to 100 (highest quality) based on higher intake of vegetables, fruit, whole grains, nuts and legumes, long-chain (n−3) fatty acids, polyunsaturated fats, and lower intake of sugar-sweetened beverages and fruit juice, red/processed meat, saturated fats, sodium (McCullough et al., 2002). BMI was calculated using participants’ self-reported height and weight (kg/m2). Self-reported weight has been shown to be highly correlated (r = .97) with weight measured by study staff within this cohort (Rimm et al., 1990).

Statistical Analysis

Statistical analyses were conducted using SAS, 9.4. We first computed the descriptive statistics for each covariate across levels of social integration, adjusting for age. A set of four accelerated failure time (AFT) models were used in primary analyses to estimate the proportion by which participants’ life spans differed in association with level of social integration (N = 72,322). Compared to the Cox proportional hazards models, the AFT models provide the advantage of easily interpretable results (i.e., percent change in life span), while still incorporating longitudinal data and controlling for multiple covariates (Swindell, 2009; Wei, 1992). AFT models were shown to be a useful statistical framework for aging research (Swindell, 2009), and have been leveraged in prior studies investigating the relationship of personality (Costa et al., 2014), optimism (Lee et al.,2019), and inflammation markers (Wassel, Barrett-Connor, & Laughlin, 2010), respectively, with longevity.

In this study, a “basic” adjusted model included potential demographic confounders (i.e., age, husband’s education, and participant’s education). A second model, the core model, further adjusted for baseline health status variables (i.e., prevalent or history of high cholesterol, high blood pressure, MI, stroke, diabetes, and cancer). A third model included both demographic confounders and health behavior-related factors (i.e., smoking status, physical activity, alcohol consumption, diet quality, and BMI) to assess whether behaviors accounted for any of the observed association between social integration and longevity. A fourth model adjusted for all covariates simultaneously. Sample size for the third and fourth models was slightly reduced as they were evaluated among women who had data on all health-behavior related variables (n = 66,684; 92.20% of the main analytic sample). We applied the transformation 100(eβ − 1) to the regression coefficient for our primary exposure, social integration, to interpret the findings as the percent change in the expected survival time comparing each social integration level to the reference (highly socially isolated). A positive coefficient suggests that greater levels of social integration are associated with greater longevity.

We conducted three additional analyses. A first sensitivity analysis considered the role of depression in a subset of women who had data on depression (n = 72,123; 99.72% of the main analytic sample) by evaluating changes in the effect estimates for social integration when including depressive symptoms in the core model controlling for demographic and health status covariates. A second sensitivity analysis evaluated the main models without excluding women who died within 2 years of baseline (n = 72,776). Finally, in secondary analyses, we examined whether any of the four domains of social integration (i.e., marriage, close friends/relatives, group associations, and religious activities) were differentially predictive of longevity. In this analysis, we evaluated separate models, considering each domain as an independent predictor in the core AFT model described earlier.

We also conducted analyses using logistic regression models to assess the likelihood of survival to the age of 85 years or older, representing exceptional longevity, using the same modeling strategy described for AFT analyses (N = 16,818 for Models 1 and 2; n = 15,598 for Models 3 and 4 [92.75% of the main analytic sample]). No standard definition for exceptional longevity has been established; however, the cut point of 85 years is commonly used (Newman & Murabito, 2013; Revelas et al., 2018) because it is well beyond the average life expectancy of individuals born in the early 20th century, without being extremely rare. Secondary analyses with the logistic regression models explored the roles of individual domains of social integration whereas sensitivity analyses assessed depressive symptoms (n = 16,757; 99.64% of the main analytic sample) as a potential confounder and the main models without excluding deaths 2 years after study onset (n = 17,016; as described earlier).

Results

Table 1 shows the age-adjusted distributions of covariates in 1992 by level of social integration for the primary analytic sample (i.e., the sample for AFT analyses). Over an average of 18.73 (SD = 4.15) years of follow-up, about a third of this sample (n = 25,723) died within the study period. Participants classified as “highly socially integrated” (highest level of social integration) had continuous SNI scores that ranged from 10 to 12, and the mean SNI score in the overall sample was 7.76. At the study baseline in 1992, 81.55% of participants were married, 46.10% reported having six or more close friends/relatives, 53.56% reported attending religious activities once a week or more, and 11.07% reported participating in group associations for 6 or more hours per week. The mean age was 58.80 years. Women categorized as highly socially integrated reported husbands having higher levels of education, were less likely to be depressed, and had healthier lifestyle (e.g., were less likely to be current or former smoker whereas more likely to be physically active).

Table 1.

Age-Adjusted Covariates by Quartiles of Social Integration in 1992 (N = 72,322)

Highly socially isolated (n = 13,481) Moderately isolated (n = 17,061) Moderately integrated (n = 22,881) Highly socially integrated (n = 18,899)
Agea 58.48 (7.12) 58.63 (7.21) 58.69 (7.10) 59.32 (7.09)
Registered nurse education level, % 69 69 72 69
Husband’s education
 -Less or equal to high school degree, % 31 36 38 35
 -More than high school degree, % 38 49 51 56
 -Missing, % 32 17 11 9
Clinical depressive symptoms, % 23 17 15 9
High cholesterol, % 45 45 47 46
High blood pressure, % 36 35 34 32
Diabetes, % 6 6 5 5
Cardiovascular disease, % 3 3 2 2
Cancer, % 10 10 10 10
Smoking status
 -Never smoker, % 33 39 46 54
 -Former smoker, % 44 44 42 39
 -Current smoker, % 24 17 12 8
≥150 min of moderate-to-vigorous physical activity/week, % 50 55 55 62
Low-to-moderate alcohol consumption, % 88 89 91 92
Body mass index 26.01 (5.70) 26.01 (5.32) 25.92 (5.13) 25.93 (5.08)
Alternative Healthy Eating Index scoreb 48.26 (10.28) 48.54 (9.91) 48.27 (9.83) 49.08 (9.55)
Open in a new tab

Note: Values are means (SD) or percentages and are standardized to the age distribution of the study population. Values of polytomous variables may not sum to 100% due to rounding.

aValue is not age adjusted.

bHigher score indicates healthier diet.

AFT analyses demonstrated a graded association between higher levels of social integration and longer life span (p value for trend ≤.0001 in all models; Table 2). In models adjusted for demographic variables and health status (core model), compared to women with the lowest levels of social integration, women who were moderately integrated and highly socially integrated had 7.06% (95% CI: 5.87–8.27) and 10.10% (95% CI: 8.80–11.42) longer life span, respectively. These associations declined to 3.74% (95% CI: 2.60–4.89) and 4.88% (95% CI: 3.66–6.11) when health behaviors were further added to the model, but remained statistically significant. In a fully adjusted model assessing SNI score as a continuous variable, each one-unit increase in social integration was modestly but significantly associated with a 0.69% (95% CI: 0.53–0.86) increase in life span.

Table 2.

Percent Change in Life Span Associated With Social Integration From 1992 to 2014 (N = 72,322)

Social integration score quartiles
Highly socially isolated Moderately isolated Moderately integrated Highly socially integrated
% 95% CI % 95% CI % 95% CI % 95% CI
Model 1: demographics 0.00 Referent 5.52 4.28–6.78 7.87 6.65–9.10 11.04 9.71–12.38
Model 2: demographics and health conditions 0.00 Referent 4.97 3.75–6.22 7.06 5.87–8.27 10.10 8.80–11.42
Model 3: demographics and health behaviorsa 0.00 Referent 2.70 1.53–3.87 3.74 2.60–4.89 4.88 3.66–6.11
Model 4: all variablesa 0.00 Referent 2.37 1.22–3.54 3.22 2.10–4.36 4.33 3.13–5.56
Open in a new tab

Note: Model 1: age, education, and husband’s education. Model 2: age, education, husband’s education, as well as prevalent/history of high cholesterol, high blood pressure, diabetes, cancer, stroke, and myocardial infarction. Model 3: age, education, husband’s education, smoking status, physical activity, alcohol, body mass index, and the Alternative Health Eating Index (AHEI) diet index. Model 4: age, education, husband’s education, as well as prevalent/history of high cholesterol, high blood pressure, diabetes, cancer, stroke, and myocardial infarction, smoking status, physical activity, alcohol, body mass index, and the AHEI. CI = confidence interval.

aSample size for these models was 66,684 because of missing data for health behaviors.

In AFT models assessing the domains of social integration separately, all four domains were associated with increased longevity in core models controlling for demographic and health status variables (Supplementary Table 2). Participants with the highest versus lowest level of participation in group associations and religious activities had a life span 3.81% longer (95% CI: 2.49–5.14) and 6.39% longer (95% CI: 5.36–7.42), respectively. Moreover, having six or more close friends versus none was related to a 7.39% (95% CI: 1.23–13.93) increase in life span, whereas being married/partnered was associated with a 6.09% increase in life span compared to being widowed/separated/divorced (95% CI: 5.05–7.13). In a sensitivity analysis assessing depressive symptoms as a potential confounder, findings were materially unchanged; for example, the effect estimate comparing the most to the least socially integrated participants declined slightly to 9.24% and remained statistically significant (95% CI: 7.94–10.55). Similarly, results were robust when including women who died within the first 2 years after study onset: in the core model, compared to women with the lowest levels of social integration, women who were moderately integrated and highly socially integrated had 8.12% (95% CI: 6.77–9.49) and 11.63% (95% CI: 10.15–13.13) longer life span, respectively.

Of the women included in analyses of exceptional longevity, 9,070 (53.93%) survived to the age of 85 years or older. Similar to findings in AFT analyses, there was a graded association between higher levels of social integration and greater likelihood of exceptional longevity (p value for trend ≤.001 in all models; Table 3). For example, in the core model, compared to highly socially isolated women, the likelihood of achieving exceptional longevity for participants who were moderately integrated and highly socially integrated was higher with odds ratios (ORs) of 1.28 (95% CI: 1.17–1.40) and 1.41 (95% CI: 1.28–1.54), respectively. These associations declined slightly, but remained statistically significant, when health behaviors were further included in the models. In a fully adjusted model assessing SNI score as a continuous variable, the OR for exceptional longevity was barely but significantly associated with a one-unit increase in social integration was 1.02 (95% CI: 1.01–1.04).

Table 3.

Odds Ratios for the Association of Social Integration With Survival Past Age of 85 (N = 16,818)

Social integration score quartiles
Highly socially isolated Moderately isolated Moderately integrated Highly socially integrated
OR 95% CI OR 95% CI OR 95% CI OR 95% CI
Model 1: demographics 1.00 Referent 1.21 1.10–1.33 1.32 1.21–1.44 1.47 1.34–1.61
Model 2: demographics and health conditions 1.00 Referent 1.19 1.08–1.31 1.28 1.17–1.40 1.41 1.28–1.54
Model 3: demographics and health behaviorsa 1.00 Referent 1.11 1.00–1.23 1.14 1.04–1.26 1.20 1.08–1.32
Model 4: all variablesa 1.00 Referent 1.09 0.98–1.21 1.11 1.01–1.22 1.15 1.04–1.27
Open in a new tab

Note: Model 1: age, education, and husband’s education. Model 2: age, education, husband’s education, as well as prevalent/history of high cholesterol, high blood pressure, diabetes, cancer, stroke, and myocardial infarction. Model 3: age, education, husband’s education, smoking status, physical activity, alcohol, body mass index, and the Alternative Health Eating Index (AHEI) diet index. Model 4: age, education, husband’s education, as well as prevalent/history of high cholesterol, high blood pressure, diabetes, cancer, stroke, and myocardial infarction, smoking status, physical activity, alcohol, body mass index, and the AHEI. CI = confidence interval; OR = odds ratio.

aSample size for these models was 15,598 because of missing data for health behaviors.

Component-specific analyses of the SNI demonstrated that greater participation in group associations, greater religious activities attendance, and currently being married/partnered were associated with greater likelihood of exceptional longevity, although there was no statistically significant association for number of close friends/relatives (Supplementary Table 3). In additional sensitivity analyses, effect estimates were materially unchanged (OR for highly socially integrated vs highly socially isolated in core model = 1.42, 95% CI: 1.30–1.56) when depressive symptoms were included in the model as a potential confounder or when women who died within the first 2 years of follow-up were not excluded (core model: ORmoderately integrated vs highly isolated = 1.29, 95% CI: 1.18–1.41; ORhighly integrated vs highly isolated = 1.43, 95% CI: 1.30–1.57).

Discussion

To the best of our knowledge, this is the largest study to assess the association between social integration and life span, and the first to consider the association between social integration and the achievement of exceptional longevity. Consistent with our hypothesis, higher levels of social integration were associated with increased life span and greater likelihood of exceptional longevity. This association persisted in models adjusting for chronic health conditions; while estimates were slightly attenuated when controlling for health behaviors, they remained statistically significant. Such attenuation is congruent with our hypothesis that health-related behaviors serve in part as pathways by which social relationships affect physical health, although further work is needed with more clear temporality between the measures of social integration and health-related behaviors and to rule out the possibility of confounding. Similarly, due to data availability, we could not definitively assess whether depression preceded or was consequent to social integration levels; however, in sensitivity analyses controlling for depressive symptoms, effect estimates barely changed and remained statistically significant. Moreover, in all analyses, CIs were fairly narrow around the effect estimates, indicating the estimates are stable. The magnitude of associations was also comparable to those observed in prior research targeting psychosocial determinants of longevity using analogous analyses (Costa et al., 2014; Lee et al., 2019).

In analyses where SNI domains were assessed separately, religious activities attendance, participation in group associations, and being married/partnered were each associated with increased life span and likelihood of exceptional longevity, whereas number of close friends and relatives was less clearly related to these outcomes. This is consistent with previous research in the same cohort that demonstrated lower risk of coronary heart disease in relation to greater social integration using the composite measure, as well as all individual components of social integration except the close friend/relatives subdomain (Chang et al., 2017). However, other research using a breast cancer patient population from this cohort demonstrated that the composite measure and the number of close friends/relatives, but not other individual domains of social integration, were associated with greater likelihood of breast cancer survival (Kroenke et al., 2006). These differences point to the potential specificity of associations between health and social relationships—what is beneficial in one set of circumstances (i.e., a healthy population) may be less effective in another (i.e., a patient population)—and lends support to the idea that associations of health to social integration may be context dependent, rather than universal. At a minimum, findings in healthy versus patient populations may not be interchangeable.

A variety of mechanisms might explain the association between greater social integration and improved health outcomes. More favorable social relationships, captured for instance by social support and social integration, are associated with healthier levels of behavioral factors, including physical activity (Kroenke et al., 2017; Tay, Tan, Diener, & Gonzalez, 2013), successful management of chronic illnesses (Gallant, 2003; Tay et al., 2013), and smoking abstinence/cessation (Kroenke et al., 2017; Wagner, Burg, & Sirois, 2004). Positive social relationships may increase likelihood of experiencing psychological well-being (e.g., optimism, positive affect) (Kubzansky et al., 2018; Steptoe, 2019), which have been associated with future engagement in health-related behaviors (Kubzansky et al., 2018; Steptoe, 2019).

Social relationships may also improve health independently of health behaviors by enhancing positive affect and feelings of belonging and self-worth, which may have direct beneficial effects on physiology through neuroendocrine and immune pathways (Berkman et al., 2000; Cohen, 1988; Kroenke, 2018). Or they may buffer potentially toxic effects of adverse experiences and psychosocial distress (Berkman & Krishna, 2014; Kubzansky et al., 2018). In both cross-sectional and longitudinal observational studies, greater social integration and other positive characteristics of social relationships have been linked to improved biomarkers of metabolic function (e.g., cholesterol and blood pressure) (Yang et al., 2016; Yang, Li, & Ji, 2013) and to reduced systemic inflammation (Penwell & Larkin, 2010; Yang et al., 2016). It is also possible that shared genetics (e.g., among biological relatives) affect both social integration and longevity. Furthermore, prior findings also showed that social relationships, including the size of one’s social network, are positively associated with cognitive abilities and slower cognitive decline in midlife and older adults (Kelly et al., 2017). Results from this study suggest health behaviors may mediate partly but not fully the social integration–longevity relationship. Thus, further research should evaluate whether biological processes and cognitive function might also be at play.

Several limitations of this study should be noted. Findings in these primarily white women may not be generalizable to minorities or to men, as both social integration and mortality rates are different in different racial/ethnic groups and in different sexes; yet, the homogeneity of this cohort enhances the study’s internal validity. As with all observational research, potential for unmeasured confounding remains possible. Nonetheless, our analyses controlled for a wide range of demographic and health status variables that may serve as confounders, including health behaviors and depressive symptoms. Because all women were married at cohort baseline (1976), most (but not all) participants were still married/partnered in 1992, hence reducing variability in the exposure at the current study baseline. However, heterogeneity increased over follow-up, as a higher proportion of women experienced separation/divorce or widowhood. Finally, although the widely studied SNI is considered a complex measure of social integration (i.e., assessing multiple dimensions), it does not capture the quality of these relationships, which also likely affect health-related outcomes (Kroenke, 2018). According to the socioemotional selectivity theory (Löckenhoff & Carstensen, 2004), individuals would progressively prioritize existing emotionally rewarding relationships over the expansion of their social network toward the end of life. As a result, it is possible that the association between larger social networks and longevity captures the fact that younger individuals who are likely to live longer, and that such bias would not be fully accounted by statistical control for chronological age. However, in the current sample, the SNI score was highly stable over 16 years and comparable across age groups, reducing concerns that these women experienced substantial changes in the size of their social network over time. These limitations are offset by several strengths, including the use of a large and well-characterized cohort, as well as a follow-up over two decades that enabled the assessment of exceptional longevity (i.e., attaining 85 years). Another strength of the study is its prospective research design, which combined with the 2-year lag introduced in our statistical analyses and control for major health conditions at baseline, reduced concerns about the potential for reverse causation.

As longer life spans become more achievable through improved disease prevention and medical technology, it is increasingly important to work toward a better understanding of psychosocial assets that can help promote longer and healthier lives. A greater appreciation of these assets may be able to inform our thinking about the resources or reserves that are necessary to help people successfully age with better health. Social integration, as demonstrated in these analyses, is one such health asset that may have a substantial association with longevity. Furthermore, although many efforts at intervention have fallen short, there is also evidence that social integration has the potential to be modifiable (Cohen & Janicki-Deverts, 2009; Holt-Lunstad et al., 2017; Kroenke, 2018). If we can develop effective ways to intervene on one’s social environment, we may be able to develop low-cost and targeted interventions to help individuals achieve longer and healthier lives.

Ethical Approval

The authors assume full responsibility for analyses and interpretation of these data. All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional review boards of the institutional and/or national research committee (Brigham and Women’s Hospital; IRB protocol number: 1999P011114) and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards. Informed consent was obtained from all individual participants included in the study.

Funding

This work was supported by the National Institutes of Health (grant number UM1 CA186107) for the Nurses’ Health Study and by a National Cancer Institute K01 Career Development Grant (grant number K01 CA169041) to Dr. R. Tucker-Seeley.

Conflict of Interest

None reported.

Supplementary Material

gbz116_suppl_Supplementary_Material
Click here for additional data file. (21.9KB, docx)

Acknowledgments

We would like to thank the participants and the staff of the Nurses’ Health Study for their valuable contributions. Further information including the procedures to obtain and access data from the Nurses’ Health Studies is described at https://www.nurseshealthstudy.org/researchers (contact email: nhsaccess@channing.harvard.edu); study materials are available at: https://www.nurseshealthstudy.org/participants/questionnaires. Analytic methods will be provided upon request to the first author. This study was not preregistered.

References

  1. Anton S. D., Woods A. J., Ashizawa T., Barb D., Buford T. W., Carter C. S., … Pahor M (2015). Successful aging: Advancing the science of physical independence in older adults. Ageing Research Reviews, 24(Pt B), 304–327. doi:10.1016/j.arr.2015.09.005 [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Bao Y., Bertoia M. L., Lenart E. B., Stampfer M. J., Willett W. C., Speizer F. E., & Chavarro J. E (2016). Origin, methods, and evolution of the three Nurses’ Health Studies. American Journal of Public Health, 106, 1573–1581. doi:10.2105/AJPH.2016.303338 [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Berkman L. F., Glass T., Brissette I., & Seeman T. E (2000). From social integration to health: Durkheim in the new millennium. Social Science and Medicine (1982), 51, 843–857. doi:10.1016/s0277-9536(00)00065-4 [DOI] [PubMed] [Google Scholar]
  4. Berkman L. F., & Krishna A (2014). 7: Social network epidemiology. In Berkman L. F., Kawachi I., & Glymour M. M. (Eds.), Social epidemiology (2nd ed, pp. 234–289). New York, NY: Oxford University Press. doi:10.1093/med/9780195377903.003.0007 [Google Scholar]
  5. Berkman L. F., & Syme S. L (1979). Social networks, host resistance, and mortality: A nine-year follow-up study of Alameda County residents. American Journal of Epidemiology, 109, 186–204. doi:10.1093/oxfordjournals.aje.a112674 [DOI] [PubMed] [Google Scholar]
  6. Chang S. C., Glymour M., Cornelis M., Walter S., Rimm E. B., Tchetgen Tchetgen E., … Kubzansky L. D (2017). Social integration and reduced risk of coronary heart disease in women: The role of lifestyle behaviors. Circulation Research, 120, 1927–1937. doi:10.1161/CIRCRESAHA.116.309443 [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Chang S. C., Pan A., Kawachi I., & Okereke O. I (2016). Risk factors for late-life depression: A prospective cohort study among older women. Preventive Medicine, 91, 144–151. doi:10.1016/j.ypmed.2016.08.014 [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Cohen S. (1988). Psychosocial models of the role of social support in the etiology of physical disease. Health Psychology, 7, 269–297. doi:10.1037/0278-6133.7.3.269 [DOI] [PubMed] [Google Scholar]
  9. Cohen S., & Janicki-Deverts D (2009). Can we improve our physical health by altering our social networks? Perspectives on Psychological Science, 4, 375–378. doi:10.1111/j.1745-6924.2009.01141.x [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Costa P. T. Jr, Weiss A., Duberstein P. R., Friedman B., & Siegler I. C (2014). Personality facets and all-cause mortality among Medicare patients aged 66 to 102 years: A follow-on study of Weiss and Costa (2005). Psychosomatic Medicine, 76, 370–378. doi:10.1097/PSY.0000000000000070 [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Gallant M. P. (2003). The influence of social support on chronic illness self-management: A review and directions for research. Health Education and Behavior, 30, 170–195. doi:10.1177/1090198102251030 [DOI] [PubMed] [Google Scholar]
  12. Holt-Lunstad J., Robles T. F., & Sbarra D. A (2017). Advancing social connection as a public health priority in the United States. American Psychologist, 72, 517–530. doi:10.1037/amp0000103 [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Kelly M. E., Duff H., Kelly S., McHugh Power J. E., Brennan S., Lawlor B. A., & Loughrey D. G (2017). The impact of social activities, social networks, social support and social relationships on the cognitive functioning of healthy older adults: A systematic review. Systematic Reviews, 6, 259. doi:10.1186/s13643-017-0632-2 [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Kroenke C. H. (2018). A conceptual model of social networks and mechanisms of cancer mortality, and potential strategies to improve survival. Translational Behavioral Medicine, 8, 629–642. doi:10.1093/tbm/ibx061 [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Kroenke C. H., Kubzansky L. D., Schernhammer E. S., Holmes M. D., & Kawachi I (2006). Social networks, social support, and survival after breast cancer diagnosis. Journal of Clinical Oncology, 24, 1105–1111. doi:10.1200/JCO.2005.04.2846 [DOI] [PubMed] [Google Scholar]
  16. Kroenke C. H., Michael Y. L., Shu X. O., Poole E. M., Kwan M. L., Nechuta S., … Chen W. Y (2017). Post-diagnosis social networks, and lifestyle and treatment factors in the after breast cancer pooling project. Psycho-Oncology, 26, 544–552. doi:10.1002/pon.4059 [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Kubzansky L. D., Huffman J. C., Boehm J. K., Hernandez R., Kim E. S., Koga H. K., … Labarthe D. R (2018). Positive psychological well-being and cardiovascular disease: JACC health promotion Series. Journal of the American College of Cardiology, 72, 1382–1396. doi:10.1016/j.jacc.2018.07.042 [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Lee L. O., James P., Zevon E., Kim E. S., Trudel-Fitzgerald C., Spiro A., … Kubzansky L. D (2019). Optimism is associated with exceptional longevity in two epidemiologic cohorts of men and women. Proceedings of the National Academy of Sciences of the United States of America, 116:18357–18362. doi:10.1073/pnas.1900712116 [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Lloyd-Jones D. M. (2014). Cardiovascular health and protection against CVD: More than the sum of the parts? Circulation, 130, 1671–1673. doi:10.1161/CIRCULATIONAHA.114.012869 [DOI] [PubMed] [Google Scholar]
  20. Löckenhoff C. E., & Carstensen L. L (2004). Socioemotional selectivity theory, aging, and health: The increasingly delicate balance between regulating emotions and making tough choices. Journal of Personality, 72, 1395–1424. doi:10.1111/j.1467-6494.2004.00301.x [DOI] [PubMed] [Google Scholar]
  21. Longo V. D., Antebi A., Bartke A., Barzilai N., Brown-Borg H. M., Caruso C., … Fontana L (2015). Interventions to slow aging in humans: Are we ready? Aging Cell, 14, 497–510. doi:10.1111/acel.12338 [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. López-Otín C., Blasco M. A., Partridge L., Serrano M., & Kroemer G (2013). The hallmarks of aging. Cell, 153, 1194–1217. doi:10.1016/j.cell.2013.05.039 [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. McCullough M. L., Feskanich D., Stampfer M. J., Giovannucci E. L., Rimm E. B., Hu F. B., … Willett W. C (2002). Diet quality and major chronic disease risk in men and women: Moving toward improved dietary guidance. American Journal of Clinical Nutrition, 76, 1261–1271. doi:10.1093/ajcn/76.6.1261 [DOI] [PubMed] [Google Scholar]
  24. National Research Council Committee on Future Directions for Behavioral and Social Sciences Research at the National Institute of Health (2001). Positive health: Resilience, recovery, primary prevention, and health promotion. Washington, DC: National Academies Press. [Google Scholar]
  25. Nausheen B., Gidron Y., Peveler R., & Moss-Morris R (2009). Social support and cancer progression: A systematic review. Journal of Psychosomatic Research, 67, 403–415. doi:10.1016/j.jpsychores.2008.12.012 [DOI] [PubMed] [Google Scholar]
  26. Newman A. B., & Murabito J. M (2013). The epidemiology of longevity and exceptional survival. Epidemiologic Reviews, 35, 181–197. doi:10.1093/epirev/mxs013 [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Penwell L. M., & Larkin K. T (2010). Social support and risk for cardiovascular disease and cancer: A qualitative review examining the role of inflammatory processes. Health Psychology Review, 4, 42–55. doi:10.1080/17437190903427546 [Google Scholar]
  28. Pinquart M., & Duberstein P. R (2010). Associations of social networks with cancer mortality: A meta-analysis. Critical Reviews in Oncology/Hematology, 75, 122–137. doi:10.1016/j.critrevonc.2009.06.003 [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Revelas M., Thalamuthu A., Oldmeadow C., Evans T. J., Armstrong N. J., Kwok J. B., … Mather K. A (2018). Review and meta-analysis of genetic polymorphisms associated with exceptional human longevity. Mechanisms of Ageing and Development, 175, 24–34. doi:10.1016/j.mad.2018.06.002 [DOI] [PubMed] [Google Scholar]
  30. Rich-Edwards J. W., Corsano K. A., & Stampfer M. J (1994). Test of the National Death Index and Equifax Nationwide Death Search. American Journal of Epidemiology, 140, 1016–1019. doi:10.1093/oxfordjournals.aje.a117191 [DOI] [PubMed] [Google Scholar]
  31. Rimm E. B., Stampfer M. J., Colditz G. A., Chute C. G., Litin L. B., & Willett W. C (1990). Validity of self-reported waist and hip circumferences in men and women. Epidemiology (Cambridge, Mass.), 1, 466–473. doi:10.1097/00001648-199011000-00009 [DOI] [PubMed] [Google Scholar]
  32. Rumpf H. J., Meyer C., Hapke U., & John U (2001). Screening for mental health: Validity of the MHI-5 using DSM-IV Axis I psychiatric disorders as gold standard. Psychiatry Research, 105, 243–253. doi:10.1016/s0165-1781(01)00329-8 [DOI] [PubMed] [Google Scholar]
  33. Steptoe A. (2019). Happiness and health. Annual Review of Public Health, 40, 339–359. doi:10.1146/annurev-publhealth-040218-044150 [DOI] [PubMed] [Google Scholar]
  34. Steptoe A., & Wardle J (2017). Life skills, wealth, health, and wellbeing in later life. Proceedings of the National Academy of Sciences of the United States of America, 114, 4354–4359. doi:10.1073/pnas.1616011114 [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Strasser T. (1978). Reflections on cardiovascular diseases. Interdisciplinary Science Reviews, 3, 225–230. doi:10.1179/030801878791925921 [Google Scholar]
  36. Swindell W. R. (2009). Accelerated failure time models provide a useful statistical framework for aging research. Experimental Gerontology, 44, 190–200. doi:10.1016/j.exger.2008.10.005 [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Tay L., Tan K., Diener E., & Gonzalez E (2013). Social relations, health behaviors, and health outcomes: A survey and synthesis. Applied Psychology. Health and Well-being, 5, 28–78. doi:10.1111/aphw.12000 [DOI] [PubMed] [Google Scholar]
  38. Trudel-Fitzgerald C., Chen Y., Singh A., Okereke O. I., & Kubzansky L. D (2016). Psychiatric, psychological, and social determinants of health in the Nurses’ Health Study cohorts. American Journal of Public Health, 106, 1644–1649. doi:10.2105/AJPH.2016.303318 [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. U.S. Department of Health and Human Services & U.S. Department of Agriculture (2010). Dietary guidelines for Americans, 2010. Washington, DC: US Government Printing Office. [Google Scholar]
  40. VanderWeele T. J., Chen Y., Long K., Kim E. S., Trudel-Fitzgerald C., & Kubzanksy L. D. (in press). Positive epidemiology? Epidemiology. [DOI] [PubMed] [Google Scholar]
  41. Wagner J., Burg M., & Sirois B (2004). Social support and the transtheoretical model: Relationship of social support to smoking cessation stage, decisional balance, process use, and temptation. Addictive Behaviors, 29, 1039–1043. doi:10.1016/j.addbeh.2004.02.058 [DOI] [PubMed] [Google Scholar]
  42. Ware J. E. Jr, & Sherbourne C. D (1992). The MOS 36-item short-form health survey (SF-36). I. Conceptual framework and item selection. Medical Care, 30, 473–483. doi:10.1097/00005650-199206000-00002 [PubMed] [Google Scholar]
  43. Wassel C. L., Barrett-Connor E., & Laughlin G. A (2010). Association of circulating C-reactive protein and interleukin-6 with longevity into the 80s and 90s: The Rancho Bernardo study. Journal of Clinical Endocrinology and Metabolism, 95, 4748–4755. doi:10.1210/jc.2010-0473 [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. Wei L. J. (1992). The accelerated failure time model: A useful alternative to the Cox regression model in survival analysis. Statistics in Medicine, 11, 1871–1879. doi:10.1002/sim.4780111409 [DOI] [PubMed] [Google Scholar]
  45. Wei J., Hou R., Zhang X., Xu H., Xie L., Chandrasekar E. K., … Goodman M (2019). The association of late-life depression with all-cause and cardiovascular mortality among community-dwelling older adults: Systematic review and meta-analysis. British Journal of Psychiatry, 215, 449–455. doi:10.1192/bjp.2019.74 [DOI] [PubMed] [Google Scholar]
  46. Woods N. F., Rillamas-Sun E., Cochrane B. B., La Croix A. Z., Seeman T. E., Tindle H. A., … Wallace R. B (2016). Aging well: Observations from the Women’s Health Initiative Study. The Journals of Gerontology, Series A: Biological Sciences and Medical Sciences, 71(Suppl. 1), S3–S12. doi:10.1093/gerona/glv054 [DOI] [PMC free article] [PubMed] [Google Scholar]
  47. Yang Y. C., Boen C., Gerken K., Li T., Schorpp K., & Harris K. M (2016). Social relationships and physiological determinants of longevity across the human life span. Proceedings of the National Academy of Sciences of the United States of America, 113, 578–583. doi:10.1073/pnas.1511085112 [DOI] [PMC free article] [PubMed] [Google Scholar]
  48. Yang Y. C., Li T., & Ji Y (2013). Impact of social integration on metabolic functions: Evidence from a nationally representative longitudinal study of US older adults. BMC Public Health, 13, 1210. doi:10.1186/1471-2458-13-1210 [DOI] [PMC free article] [PubMed] [Google Scholar]
  49. Yuan C., Spiegelman D., Rimm E. B., Rosner B. A., Stampfer M. J., Barnett J. B., … Willett W. C (2017). Validity of a dietary questionnaire assessed by comparison with multiple weighed dietary records or 24-Hour recalls. American Journal of Epidemiology, 185, 570–584. doi:10.1093/aje/kww104 [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

gbz116_suppl_Supplementary_Material
Click here for additional data file. (21.9KB, docx)

Articles from The Journals of Gerontology Series B: Psychological Sciences and Social Sciences are provided here courtesy of Oxford University Press

RESOURCES