Lisa F. Berkman

From social integration to health: Durkheim in the new millennium☆

It is widely recognized that social relationships and affiliation have powerful effects on physical and mental health. When investigators write about the impact of social relationships on health, many terms are used loosely and interchangeably including social networks, social ties and social integration. The aim of this paper is to clarify these terms using a single framework. We discuss: (1) theoretical orientations from diverse disciplines which we believe are fundamental to advancing research in this area; (2) a set of definitions accompanied by major assessment tools; and (3) an overarching model which integrates multilevel phenomena. Theoretical orientations that we draw upon were developed by Durkheim whose work on social integration and suicide are seminal and John Bowlby, a psychiatrist who developed attachment theory in relation to child development and contemporary social network theorists. We present a conceptual model of how social networks impact health. We envision a cascading causal process beginning with the macro-social to psychobiological processes that are dynamically linked together to form the processes by which social integration effects health. We start by embedding social networks in a larger social and cultural context in which upstream forces are seen to condition network structure. Serious consideration of the larger macro-social context in which networks form and are sustained has been lacking in all but a small number of studies and is almost completely absent in studies of social network influences on health. We then move downstream to understand the influences network structure and function have on social and interpersonal behavior. We argue that networks operate at the behavioral level through four primary pathways: (1) provision of social support; (2) social influence; (3) on social engagement and attachment; and (4) access to resources and material goods.

Social Ties and Mental Health

It is generally agreed that social ties play a beneficial role in the maintenance of psychological well-being. In this targeted review, we highlight four sets of insights that emerge from the literature on social ties and mental health outcomes (defined as stress reactions, psychological well-being, and psychological distress, including depressive symptoms and anxiety). First, the pathways by which social networks and social supports influence mental health can be described by two alternative (although not mutually exclusive) causal models—the main effect model and the stress-buffering model. Second, the protective effects of social ties on mental health are not uniform across groups in society. Gender differences in support derived from social network participation may partly account for the higher prevalence of psychological distress among women compared to men. Social connections may paradoxically increase levels of mental illness symptoms among women with low resources, especially if such connections entail role strain associated with obligations to provide social support to others. Third, egocentric networks are nested within a broader structure of social relationships. The notion of social capital embraces the embeddedness of individual social ties within the broader social structure. Fourth, despite some successes reported in social support interventions to enhance mental health, further work is needed to deepen our understanding of the design, timing, and dose of interventions that work, as well as the characteristics of individuals who benefit the most.

Two Shorter Forms of the CES-D Depression Symptoms Index

Brief measurement devices can alleviate respondent burden and lower refusal rates in surveys. This article reports on a field test of two shorter forms of the Center for Epidemiological Studies Depression (CES-D) symptoms index in a multisite survey of persons 65 and older. Factor analyses demonstrate that the briefer forms tap the same symptom dimensions as does the original CES-D, and reliability statistics indicate that they sacrifice little precision. Simple transformations are presented to show how scores from the briefer forms can be compared to those of the original.

Genetic Susceptibility to Death from Coronary Heart Disease in a Study of Twins

BACKGROUND A family history of premature coronary heart disease has long been thought to be a risk factor for coronary heart disease. Using data from 26 years of follow-up of 21,004 Swedish twins born between 1886 and 1925, we investigated this issue further by assessing the risk of death from coronary heart disease in pairs of monozygotic and dizygotic twins. METHODS The study population consisted of 3298 monozygotic and 5964 dizygotic male twins and 4012 monozygotic and 7730 dizygotic female twins. The age at which one twin died of coronary heart disease was used as the primary independent variable to predict the risk of death from coronary heart disease in the other twin. Information about other risk factors was obtained from questionnaires administered in 1961 and 1963. Actuarial life-table analysis was used to estimate the cumulative probability of death from coronary heart disease. Relative-hazard estimates were obtained from a multivariate survival analysis. RESULTS Among the men, the relative hazard of death from coronary heart disease when ones twin died of coronary heart disease before the age of 55 years, as compared with the hazard when ones twin did not die before 55, was 8.1 (95 percent confidence interval, 2.7 to 24.5) for monozygotic twins and 3.8 (1.4 to 10.5) for dizygotic twins. Among the women, when ones twin died of coronary heart disease before the age of 65 years, the relative hazard was 15.0 (95 percent confidence interval, 7.1 to 31.9) for monozygotic twins and 2.6 (1.0 to 7.1) for dizygotic twins. Among both the men and the women, whether monozygotic or dizygotic twins, the magnitude of the relative hazard decreased as the age at which ones twin died of coronary heart disease increased. The ratio of the relative-hazard estimate for the monozygotic twins to the estimate for the dizygotic twins approached 1 with increasing age. These relative hazards were little influenced by other risk factors for coronary heart disease. CONCLUSIONS Our findings suggest that at younger ages, death from coronary heart disease is influenced by genetic factors in both women and men. The results also imply that the genetic effect decreases at older ages.

The Role of Social Relations in Health Promotion

In considering new paradigms for the prevention and treatment of disease and disability, we need to incorporate ways to promote social support and develop family and community strengths and abilities into our interventions. There is now a substantial body of evidence that indicates that the extent to which social relationships are strong and supportive is related to the health of individuals who live within such social contexts. A review of population-based research on mortality risk over the last 20 years indicates that people who are isolated are at increased mortality risk from a number of causes. More recent studies indicate that social support is particularly related to survival postmyocardial infarction. The pathways that lead from such socioenvironmental exposures to poor health outcomes are likely to be multiple and include behavioral mechanisms and more direct physiologic pathways related to neuroendocrine or immunologic function. For social support to be health promoting, it must provide both a sense of belonging and intimacy and must help people to be more competent and self-efficacious. Acknowledging that health promotion rests on the shoulders not only of individuals but also of their families and communities means that we must commit resources over the next decade to designing, testing, and implementing interventions in this area.

Social Disengagement and Incident Cognitive Decline in Community-Dwelling Elderly Persons

Social engagement, which is defined as the maintenance of many social connections and a high level of participation in social activities, has been thought to prevent cognitive decline in elderly persons. Associations between a socially engaged lifestyle and higher scores on memory and intelligence tests have been observed among community-dwelling older persons (1-5). Short-term interventions to foster social and intellectual engagement have enhanced cognition among nursing home residents (6) and patients with dementia (7). In animal studies (8), mature rodents exposed to complex social and inanimate environments showed better maze-learning ability than those in sparser surroundings. Social engagement challenges persons to communicate effectively and participate in complex interpersonal exchanges. Besides providing a dynamic environment that requires the mobilization of cognitive faculties, social engagement may also indicate a commitment to community and family and engender a health-promoting sense of purpose and fulfillment. Another putative benefit of social engagement is greater availability of emotional support from relatives and friends. Lack of such support can predict adverse health outcomes (9), but its influence on cognitive decline has not been examined. Although published findings on cognitive function are suggestive, interpretation of available epidemiologic data is hampered by methodologic and conceptual shortcomings. Most studies are not longitudinal assessments of representative population-based cohorts but cross-sectional observations of volunteers or other special samples. Adjustment for potential confounders, such as education or health status, is often lacking. Finally, although the concept of social engagement is intuitively accessible, it has been difficult to measure this construct. Early studies viewed previous favorable socioeconomic or occupational status as synonymous with social engagement in old age (1); this precluded examination of more salient or malleable postretirement behaviors. Recent investigations (10-12), however, have used narrowly defined indicators (for example, marital status or specific recreational activities) to make inferences about the effect of late-life social environments on cognition. Other investigations (5) have used composite activity scales as proxies for social engagement but have not distinguished between activities that demand a high level of externally directed attention and those that do not. Social and solitary pursuits are also rarely disaggregated, although each may affect cognitive function differently. We determined whether a global measure of social disengagement was associated with incident cognitive impairment in a large cohort of community-dwelling elderly persons followed for 12 years. We acknowledge the potential salutary influence of solitary mental pursuits, but we focused on whether interpersonal connections and activities can prevent cognitive decline. Convincing demonstration of such an association could motivate elderly persons, their families, and their care providers to maintain active relationships. It would also justify continued funding for community-based and institutionally based social programs for elderly persons and the adoption of social policies in which activities of older persons are valued. Methods Participants The study sample was drawn from the New Haven, Connecticut, site of the Established Populations for Epidemiologic Studies of the Elderly (EPESE) project, described in detail elsewhere (13). The New Haven cohort is a multistage probability sample of 2812 noninstitutionalized persons 65 years of age or older who were living in New Haven, Connecticut, in 1982. Samples were drawn from three housing strata: public (income-restricted) housing for elderly persons, private housing for elderly persons, and community housing. Women were randomly subsampled to achieve equal representation of both sexes. The baseline response rate was 82%. Trained lay examiners interviewed members of the cohort at home in 1982, 1985,1988, and 1994 and by telephone in intervening years. Our study was approved by the institutional review board of Yale University. Measurements Cognitive Function Cognitive performance was measured during in-home interviews with the 10-item Short Portable Mental Status Questionnaire (SPMSQ) (14). (The original item What is the name of this place? was changed to What is your address? because this seemed more appropriate for community-dwelling persons.) Correct answers received 1 point each. If respondents declined to answer 4 or more items or if answers to 4 or more items were missing, the questionnaire was not scored. Otherwise, a participants decision not to answer a question was scored as an incorrect answer and scores on missing items were imputed by assigning the mean score of the nonmissing items. As has been done previously (15-17), scores were divided into three categories: high (a score of 9 or 10), medium (a score of 7 or 8), and low (a score of 0 to 6). The validity of the questionnaire as a cognitive impairment measure has been assessed in a subsample of respondents at another EPESE site. Participants received detailed medical examinations to determine the presence and severity of cognitive impairment. When medium scores on the SPMSQ and the category of mild impairment were excluded, the questionnaires sensitivity and specificity in identifying moderate or severe impairment were 85% and 96%, respectively (18). Social Disengagement A comprehensive assessment of social connections and activities was completed during in-home interviews. We examined six indicators of social engagement: presence of a spouse, monthly visual contact with at least three relatives or close friends, yearly nonvisual contact (telephone calls or letters) with at least 10 relatives or close friends, frequent attendance (at least once per month) at religious services, membership in other groups, and regular participation in recreational social activities. The Appendix provides exact questions and coding rules for these indicators. A composite social disengagement index was constructed from the six indicators. Five to six social ties received a score of 1, three to four social ties received a score of 2, one to two social ties received a score of 3, and no social ties received a score of 4. Tie refers to any type of social contact. If scores for more than two indicators were missing, the index was not scored. Emotional Support Perceived availability of emotional support from a social network was assessed with the question, Can you count on anyone to provide you with emotional supportthat is, talking over problems or helping make a difficult decision? The adequacy of this support was assessed by the question, Could you have used more emotional support than you received? Covariates We selected the following self-reported variables as potential confounders because they had cross-sectional associations with disengagement or impaired cognition among the cohort or because they are established risk factors for Alzheimer disease or vascular dementia (both of which are prevalent causes of progressive cognitive decline in elderly persons) (19, 20). Sociodemographic factors were age, sex, ethnicity (white or nonwhite), education ( 12 years or<12 years), annual income (<

Emotional support and survival after myocardial infarction. A prospective, population-based study of the elderly.

10 000,

Driving Cessation and Increased Depressive Symptoms: Prospective Evidence from the New Haven EPESE

10 000, or not given), and housing. Health status indicators were the presence of physical disability, which was defined as requiring assistance from another person or special equipment with at least one activity of daily living (walking across a room, dressing, eating, transferring from bed to chair, bathing, or using the toilet) (21) or limitations in gross mobility (climbing one flight of stairs or walking half a mile) (22); high-risk cardiovascular profile (compared with low risk, where high risk was defined as a measured sitting blood pressure>160/95 mm Hg or a history of physician-diagnosed stroke, diabetes, or myocardial infarction); visual impairment (difficulty in reading ordinary newspaper print); auditory impairment (difficulty in hearing what a person says without seeing his face if that person talks in a normal voice in a quiet room); and symptoms of depression (a score 16 on the Center for Epidemiologic Studies Depression Scale [23]). Health-related behaviors were current smoking status, alcohol consumption, and level of physical activity (described in the Appendix). Statistical Analysis Cognitive decline was defined as a transition to a lower SPMSQ category (transition from high to medium or low or transition from medium to low) during a given interval. Respondents with low SPMSQ scores at the beginning of an interval were excluded from consideration during that interval. Incidence of cognitive decline by initial level of social disengagement was estimated over intervals of three lengths:3-year intervals (1982 to 1985 and 1985 to 1988), 6-year intervals (1982 to1988 and 1988 to 1994), and a 12-year interval (1982 to 1994). Approximately20% of respondents with low SPMSQ scores in 1982 scored higher in 1985 or 1988and were therefore included in the 1985 to 1988 or 1988 to 1994 analyses. Polytomous logistic regression was used to estimate the relative risk for cognitive decline or death by level of disengagement, controlling for potential confounders. The three outcomesmaintenance of cognitive function, cognitive decline, and deathwere treated as unordered categorical variables. To examine the validity of the assumption of a linear dose-response relation between social disengagement and cognitive decline or death, disengagement was initially coded as a set of indicator variables, each corresponding to 1 point on the composite disengagement index. Inspection of the resulting coefficients suggested that modeling disengagement as an ordered categorical covariate was justified; the addition of quadratic or cubic terms did not substantially improve model fit.

Social relationships, social support, and patterns of cognitive aging in healthy, high-functioning older adults: MacArthur Studies of Successful Aging.

OBJECTIVE To compare the survival of elderly patients hospitalized for acute myocardial infarction who have emotional support with that of patients who lack such support, while controlling for severity of disease, comorbidity, and functional status. DESIGN A prospective, community-based cohort study. SETTING Two hospitals in New Haven, Connecticut. PATIENTS Men (n = 100) and women (n = 94) 65 years of age or more hospitalized for acute myocardial infarction between 1982 and 1988. MEASUREMENTS Social support, age, gender, race, education, marital status, living arrangements, presence of depression, smoking history, weight, and physical function were assessed prospectively using questionnaires. The presence of congestive heart failure, pulmonary edema, and cardiogenic shock; the position of infarction; in-hospital complications; and history of myocardial infarction were assessed using medical records. Comorbidity was defined using an index based on the presence of eight conditions. RESULTS Of 194 patients, 76 (39%) died in the first 6 months after myocardial infarction. In multiple logistic regression analyses, lack of emotional support was significantly associated with 6-month mortality (odds ratio, 2.9; 95% CI, 1.2 to 6.9) after controlling for severity of myocardial infarction, comorbidity, risk factors such as smoking and hypertension, and sociodemographic factors. CONCLUSIONS When emotional support was assessed before myocardial infarction, it was independently related to risk for death in the subsequent 6 months.

Predictors of cognitive change in older persons : MacArthur studies of successful aging

OBJECTIVES: The purpose of this study was to determine the association between driving cessation and depressive symptoms among older drivers. Previous efforts in this area have focused on the factors associated with cessation, not the consequences of having stopped.