Social Network Study of Health Effects in Aging

Harvard Medical School, Boston MA

Investigators

Linked publications & trials

Abstract

The health of humans who are linked through a social tie may be inter-dependent. Illness, disability, health behaviors, health care use, or death in one person may be associated with similar outcomes in specific others to whom that person is connected. Such inter-individual health effects, operating through a diverse set of mechanisms, may obtain in social relations ranging from spouses to siblings, parents, friends, co-workers, or neighbors. We propose to examine the existence, size, and range of such effects in the context of cardiovascular disease, which is annually responsible for 40% of deaths in the U.S. and $350 billion in medical costs and lost productivity. To do so, we will use the landmark Framingham Heart Study supplemented by heretofore unused administrative data. We have four specific aims. First, we will develop a multi-purpose panel data set containing demographic, social, behavioral, clinical, morbidity, and mortality information about a cohort of 5,124 people and also their identified social relations, followed for 30 years. Each individual in our sample will have an average of more than five other individuals of diverse relationships in our sample to whom they are connected. Second, we will describe the attributes of each individual's social contacts and examine how such attributes affect the onset and outcome of cardiovascular disease; for example, we will examine whether having a more educated social network is salubrious. Third, we will, as our key aim, evaluate how health behaviors in a person embedded in a social network depend on prior health events in others in their social network. Our specific test case is whether a heart attack or stroke in one's social contacts is associated with subsequent weight loss, smoking cessation, or aspirin use. We will evaluate how these effects vary according to the nature of the relationship - that is, depending on whether the contact is a spouse, parent, sibling, friend, co-worker, or neighbor. And we will evaluate whether these effects vary depending on baseline social and clinical traits of the subject (e.g., whether the subject is obese, diabetic, uneducated, etc.) or the contact (e.g., whether the contact is young, etc.). Fourth, we will evaluate whether disability (e.g., due to a heart attack or stroke) in contacts is associated with subsequent disability in subjects. Our work has implications for the understanding of cardiovascular risk behaviors and outcomes, for the understanding of social networks, for the demography of aging, and for policy concerns as diverse as the role of neighborhoods in health to the optimal estimation of cost-efectiveness of medical care and behavioral interventions.