Claire Yang

Social relationships and physiological determinants of longevity across the human life span.

Significance Although much evidence has accrued in research over the past 20 years on the strong causal associations between social relationships and health and longevity, important gaps remain in our understanding of the mechanisms, timing, and duration of these associations. This study integrates social and biological disciplinary perspectives and research to examine how social relationships “get under the skin” to affect physiological well-being as individuals age. By combining data from and harmonizing measurement across four large nationally representative, population-based, contemporary surveys using an innovative longitudinal life course design, this study provides previously unidentified evidence on the biological and life course mechanisms linking social relationship patterns with health. As such, our findings advance explanations of the emergence and progression of diseases across the human life span. Two decades of research indicate causal associations between social relationships and mortality, but important questions remain as to how social relationships affect health, when effects emerge, and how long they last. Drawing on data from four nationally representative longitudinal samples of the US population, we implemented an innovative life course design to assess the prospective association of both structural and functional dimensions of social relationships (social integration, social support, and social strain) with objectively measured biomarkers of physical health (C-reactive protein, systolic and diastolic blood pressure, waist circumference, and body mass index) within each life stage, including adolescence and young, middle, and late adulthood, and compare such associations across life stages. We found that a higher degree of social integration was associated with lower risk of physiological dysregulation in a dose–response manner in both early and later life. Conversely, lack of social connections was associated with vastly elevated risk in specific life stages. For example, social isolation increased the risk of inflammation by the same magnitude as physical inactivity in adolescence, and the effect of social isolation on hypertension exceeded that of clinical risk factors such as diabetes in old age. Analyses of multiple dimensions of social relationships within multiple samples across the life course produced consistent and robust associations with health. Physiological impacts of structural and functional dimensions of social relationships emerge uniquely in adolescence and midlife and persist into old age.

Social support, social strain and inflammation: Evidence from a national longitudinal study of U.S. adults

Social relationships have long been held to have powerful effects on health and survival, but it remains unclear whether such associations differ by function and domain of relationships over time and what biophysiological mechanisms underlie these links. This study addressed these gaps by examining the longitudinal associations of persistent relationship quality across a ten year span with a major indicator of immune function. Specifically, we examined how perceived social support and social strain from relationships with family, friends, and spouse at a prior point in time are associated with subsequent risks of inflammation, as assessed by overall inflammation burden comprised of five markers (C-reactive protein, interleukin-6, fibrinogen, E-selectin, and intracellular adhesion molecule-1) in a national longitudinal study of 647 adults from the Midlife Development in the United States (1995-2009). Results from multivariate regression analysis show that (1) support from family, friends, and spouse modestly protected against risks of inflammation; (2) family, friend, and total social strain substantially increased risks of inflammation; and (3) the negative associations of social strain were stronger than the positive associations of social support with inflammation. The findings highlight the importance of enriched conceptualizations, measures, and longitudinal analyses of both social and biological stress processes to elucidate the complex pathways linking social relationships to health and illness.

Social Relationships and Hypertension in Late Life Evidence From a Nationally Representative Longitudinal Study of Older Adults

Objective: Social relationships are widely understood to be important for sustaining and improving health and longevity, but it remains unclear how different dimensions of social relationships operate through similar or distinct mechanisms to affect biophysiological markers of aging-related disease over time. Method: This study utilized longitudinal data on a nationally representative sample of older adults from the National Social Life, Health, and Aging Project (2005-2011) to examine the prospective associations between social integration and social support and change in systolic blood pressure (SBP) and hypertension risk over time. Results: Although both social relationship dimensions have significant physiological impacts, their relative importance differs by outcome. Low social support was predictive of increase in SBP, whereas low social integration was predictive of increase in risk of hypertension. Discussion: The different roles of relationship characteristics in predicting change in physiological outcomes suggest specific biophysiological stress response and behavioral mechanisms that have important implications for both scientific understandings and effective prevention and control of a leading chronic condition in late life.

Misunderstandings, Mischaracterizations, and the Problematic Choice of a Specific Instance in Which the IE Should Never Be Applied

It is our pleasure to respond to the Editors request for a commentary on Luos article (this issue), which critically evaluates the utility of the intrinsic estimator (IE) that we first introduced to demography and sociology in Yang et al. (2004). We first respond to the Editors request for “an assessment of the authors argument, evidence, and conclusions.” It is truly unfortunate and fundamentally incorrect to interpret our stance on the IE method as seeing it a “holy grail” or “magic bullet” for the identification problem of the age-period-cohort (APC) accounting model/multiple classification model. Nowhere in our previous publications did we make such a claim. We were crystal clear about the circumstances in which a full-blown APC models should be used. And such circumstances equally apply to any estimator of full three-factor APC models, not just the IE. Luo completely lost sight of this starting point. Works as early as Yang (2008) and as recent as Yang and Land (2013: chapter 5) have laid out a three-step procedure that should be thoroughly applied to APC analysis using the accounting model. It is so important that we believe it is worth repeating here. Step 1 is to conduct descriptive data analyses using graphics, with the objective being to provide qualitative understanding of patterns of temporal variations. Step 2 is model fitting and calculation of model fit statistics, such as the Bayesian information criterion (BIC). The objective is to ascertain whether the data are sufficiently well described by any single factor or two-factor model of age (A), time period (P), and cohort (C) effects for which there is no identification problem. Only when these analyses suggest that all three dimensions are operative should one proceed with Step 3: a three-factor APC model to which a constrained estimator can be applied to identify the A, P, and C effects. By revisiting Glenns (2005) numerical example, Yang and Land (2013:109) emphasized that “imposition of a full APC model on data when a reduced model fits the data equally well or better constitutes a model misspecification and should be avoided.” Empirical examples of chronic disease mortality in Yang (2008) and cancer mortality in Yang and Land (2013) showed the necessity of all three steps, whereas those of cancer incidence for certain sites in the latter show that the first two steps suffice. A blind application of the IE, or any other constrained estimator, of the full three-factor APC model was never recommended. It follows that any of Luos exercises based on scenarios in which no full APC models should be applied are invalid. Because space limitations imposed on this comment do not permit a more detailed analysis of Luos article , we posted a full response elsewhere (http://www.unc.edu/~yangy819/apc/index.html) that explains the fundamental flaws. Briefly, Luo claims that (1) there is nothing new about the IE for the age, period, and cohort effect coefficients of the classical APC accounting model; (2) the IE is not an unbiased estimator of the unidentified coefficient vector of this model; (3) the constraint imposed by the IE on the unidentified coefficient is “implicit”; and (4) the IE performs poorly as a statistical estimator of the unidentified coefficient vector when that vector has very large effects of the design matrix. In our more detailed analysis, we respond that point 1 represents a misunderstanding of the IE, point 2 is a claim that we never made, point 3 is a mischaracterization of our work, and point 4 disregards the asymptotic properties of the IE. In short, there is little merit in Luos article other than an algebraic demonstration of a situation—identical linear or nonlinear algebraic trends in the effect coefficients for all three temporal dimensions—in which the IE should never be applied.1 We next respond to the Editors suggestion to comment on the implications of the article for previous research that has used the IE approach. Given the aforementioned fallacies in Luos article, it should have no bearing on previous research that has used the IE approach. More importantly, since the publications of our earlier works on the IE, research has found increasing support for the utility of the IE method when the full APC accounting model can be properly applied. Additional methodological studies of its properties have continued. For example, Powers (2013:1039) couched the IE in a longer statistical tradition of estimable functions, showing “a number of equivalent numerical methods (i.e., principal components, singular value decomposition, generalized inverse, and so on) may be employed in various ways to provide solutions.” And Powers’ (2012) work provides a flexible extension (ie_rate) of Schulhofer-Wohl and Yangs (2006) APC IE Stata program (apc_ie). In addition, model validation studies have further demonstrated the robustness of the statistical properties of the IE through comparisons of results from an IE analysis of empirical data with results from an analysis of the same data by application of a different family of models that do not use the same identifying constraint. There is strong evidence that validates the IE method across multiple data sets and outcomes. For example, Masters et al. (2013) found estimates of temporal trends in official U.S. adult mortality rates from models using the IE to be entirely consistent with coefficient estimates from Bayesian hierarchical models using Gibbs sampling (Yang 2006). Other examples of consistent A, P, and C patterns estimated by the IE as well as alternative models such as mixed-effects models include verbal test scores from the General Social Survey (Yang et al. 2008), adult mortality from vital statistics (Yang 2008) and National Health and Interview Surveys (Masters 2012; Masters et al. 2012), and cancer incidence and mortality from national tumor registries (Yang and Land 2013). Finally, the Editor asked that we share advice about the options for future researchers who seek to identify separate age, period, and cohort effects. Our advice is that researchers should never naively apply any estimator to APC data and expect to obtain meaningful results. In all cases, APC analysis should be approached with great caution and awareness of its many pitfalls. In this context, the IE has been shown to be a useful approach to the identification and estimation of the three-factor APC accounting model. The IE, however, is not a “final” or “universal” solution to the identification problem of linear or APC accounting models. There will never be such a solution within the confines of conventional linear models that necessarily beget the identification problem.

Mortality Increase in Late-Middle and Early-Old Age: Heterogeneity in Death Processes as a New Explanation

Deviations from the Gompertz law of exponential mortality increases in late-middle and early-old age are commonly neglected in overall mortality analyses. In this study, we examined mortality increase patterns between ages 40 and 85 in 16 low-mortality countries and demonstrated sex differences in these patterns, which also changed across period and cohort. These results suggest that the interaction between aging and death is more complicated than what is usually assumed from the Gompertz law and also challenge existing biodemographic hypotheses about the origin and mechanisms of sex differences in mortality. We propose a two-mortality model that explains these patterns as the change in the composition of intrinsic and extrinsic death rates with age. We show that the age pattern of overall mortality and the population heterogeneity therein are possibly generated by multiple dynamics specified by a two-mortality model instead of a uniform process throughout most adult ages.

Social Network Ties and Inflammation in U.S. Adults with Cancer

The growing evidence linking social connectedness and chronic diseases such as cancer calls for a better understanding of the underlying biophysiological mechanisms. This study assessed the associations between social network ties and multiple measures of inflammation in a nationally representative sample of adults with a history of cancer (N = 1,075) from the National Health and Nutrition Examination Survey III (1988–94). Individuals with lower social network index (SNI) scores showed significantly greater inflammation marked by C-reactive protein and fibrinogen, adjusting for age and sex. Compared to fully socially integrated individuals (SNI = 4), those who were more socially isolated or had a SNI score of 3 or less exhibited increasingly elevated inflammation burdens. Specifically, the age- and sex-adjusted odds ratios (95%CI) for SNIs of 3, 2, and 0–1 were 1.49 (1.08, 2.06), 1.69 (1.21, 2.36), and 2.35 (1.62, 3.40), respectively (p < .001). Adjusting for other covariates attenuated these associations. The SNI gradients in the risks of inflammation were particularly salient for the lower socioeconomic status groups and remained significant after adjusting for other social, health behavioral, and illness factors. This study provided initial insights into the immunological pathways by which social connections are related to morbidity and mortality outcomes of cancer in particular and aging-related diseases in general.

Early-Life Socioeconomic Status and Adult Physiological Functioning: A Life Course Examination of Biosocial Mechanisms

ABSTRACT A growing literature has demonstrated a link between early-life socioeconomic conditions and adult health at a singular point in life. No research exists, however, that specifies the life course patterns of socioeconomic status (SES) in relation to the underlying biological processes that determine health. Using an innovative life course research design consisting of four nationally representative longitudinal datasets that collectively cover the human life span from early adolescence to old age (Add Health, MIDUS, NSHAP, and HRS), we address this scientific gap and assess how SES pathways from childhood into adulthood are associated with biophysiological outcomes in different adult life stages. For each dataset, we constructed standardized composite measures of early-life SES and adult SES and harmonized biophysiological measurements of immune and metabolic functioning. We found that the relative importance of early-life SES and adult SES varied across young, mid, and late adulthood, such that early-life SES sets a life course trajectory of socioeconomic well-being and operates through adult SES to influence health as adults age. We also documented evidence of the detrimental health effects of downward mobility and persistent socioeconomic disadvantage. These findings are the first to specify the life course patterns of SES that matter for underlying biophysiological functioning in different stages of adulthood. The study thus contributes new knowledge critical for improving population health by identifying the particular points in the life course at which interventions might be most effective in preventing disease and premature mortality.

Young Adult Risk Factors for Cancer: Obesity, Inflammation, and Sociobehavioral Mechanisms

INTRODUCTION The paper assesses social disparities in the burdens of metabolic and inflammatory risks for cancer in the U.S. young adult population and examines psychosocial and behavioral mechanisms in such disparities. METHODS Using data of 7,889 individuals aged 12-32 years from the National Longitudinal Study of Adolescent to Adult Health from 1994 to 2009, generalized linear models were used to assess the sex, race/ethnicity, and SES differences in the risks of obesity and inflammation, measured by C-reactive protein. Further tests examined the extent to which social isolation, smoking, physical inactivity, alcohol abuse, and illicit drug use explain social differentials in each biomarker outcome. RESULTS Women, blacks, Hispanics, and socioeconomically disadvantaged groups had higher risks of obesity and elevated C-reactive protein, with the SES gradients being more pronounced in female participants. Health-related behaviors showed large variation across sex, race, and SES strata. After adjusting for these behavioral variables, sex, and race disparities in obesity and excess inflammation in blacks diminished, whereas the adolescent SES disparity in obesity remained. The associations of adolescent and young adult SES disadvantage and inflammation were also explained by behavioral mechanisms. Behavioral factors associated with higher risks of obesity and inflammation differed, with the exception of fast food consumption, a risk factor for both. CONCLUSIONS This study provides new knowledge of social distribution of early life exposures to physiologic precedents to cancer development later in life with implications for prevention and early intervention of modifiable risky behaviors in adolescents and young adults.

Socioeconomic Status and Biological Risks for Health and Illness Across the Life Course

Objectives We assess the temporal properties and biosocial mechanisms underlying the associations between early-life SES and later health. Using a life course design spanning adolescence to older adulthood, we assess how early-life and various dimensions of adult SES are associated with immune and metabolic function in different life stages and examine possible bio-behavioral and psychosocial mechanisms underlying these associations. Method Data for this study come from three national studies that collectively cover multiple stages of the life course (Add Health, MIDUS, and HRS). We estimated generalized linear models to examine the prospective associations between early-life SES, adult SES, and biomarkers of chronic inflammation and metabolic disorder assessed at follow-up. We further conducted formal tests of mediation to assess the role of adult SES in linking early SES to biological functions. Results We found that early-life SES exerted consistent protective effects for metabolic disorder across the life span, but waned with time for CRP. The protective effect of respondent education remained persistent for CRP but declined with age for metabolic disorder. Adult income and assets primarily protected respondents against physiological dysregulation in middle and old ages, but not in early adulthood. Discussion These findings are the first to elucidate the life course patterns of SES that matter for underlying physiological functioning during the aging process to produce social gradients in health.

Social Relationships, Inflammation, and Cancer Survival

Background: Social stressors, such as social relationship deficits, have been increasingly linked to chronic disease outcomes, including cancer. However, critical gaps exist in our understanding of the nature and strength of such links, as well as the underlying biological mechanisms relating social relationships to cancer progression and survival. Methods: Utilizing novel questionnaire and biomarker data from the UNC Health Registry/Cancer Survivorship Cohort, this study examines the associations between diverse measures of social support and mortality risk among individuals with cancer (N = 1,004). We further assess the role of multiple serum markers of inflammation, including high-sensitivity C-reactive protein (CRP), IL6, TNFα, and VEGF, as potential mediators in the social relationship–cancer link. Results: The findings revealed that ones appraisal of their social support was associated with cancer mortality, such that individuals reporting higher levels of social support satisfaction had lower mortality risk than individuals reporting lower levels of satisfaction. The amount of support received, on the other hand, was not predictive of cancer survival. We further found evidence that inflammatory processes may undergird the link between social support satisfaction and mortality among individuals with cancer, with individuals reporting higher levels of social support satisfaction having lower levels of CRP, IL6, and TNFα. Conclusions: These results provide new knowledge of the biosocial processes producing population disparities in cancer outcomes. Impact: Our study offers new insights for intervention efforts aimed at promoting social connectedness as a means for improving cancer survival. Cancer Epidemiol Biomarkers Prev; 27(5); 541–9. ©2018 AACR.