Category: Social and Community Context

24 - Effects of Social Networks on Diabetes Outcomes Among Midlife and Older Americans

Introduction: Type II diabetes is a chronic, costly health condition and the seventh leading cause of death in the U.S. With the increasing aging population and obesity epidemic, prevalence of diabetes is projected to increase, resulting in substantial person and healthcare costs in the near future. It is imperative to target modifiable risk factors to prevent or reverse diabetes, with a particular focus on the social context of people living with diabetes. While most research is focused on social support function of social network, social capital in relation to diabetes is less understood.

Objective: Guided by Burt’s theory of social capital, the present study will cross-sectionally investigate a potential pathway through which social networks may influence the diabetes risk in community-dwelling women and men.

A U.S. nationally-representative, panel sample of 2,249 adults from the National Social Life, Health, and Aging Project (NSHAP) multistage area probability cohort is used to test all hypotheses. African-Americans, Latinos, men and individuals 75-84 were oversampled. Diabetes is ascertained by: (1) a dichotomized self-report question “Has a medical doctor ever told you that you have diabetes or high blood sugar?” and (2) continuous measure of HbA1c level. Social capital is assessed using 4 social network structures: network degree, effective size, efficiency, and constraint. Using generalized linear mixed model, we will estimate the relationship between each social network property with diabetes outcomes. A moderating effect of education will also be explored. Directed acyclic graph (DAG) will be utilized to reduce confounding and selection bias.

The sample is comprised of adults with average weighted mean of 67 years, 52% women, and 19% of respondents identified with non-white racial/ethnic group. Majority of the sample (73%) had less than a bachelor’s degree, with 16% reported having less than high school diploma.  Seventy-two percent indicated being married or co-habiting with a partner. Almost eighteen percent reported being diagnosed with diabetes. NOTE: Main analyses are still being conducted and unavailable at this time.


Ruslan V. Nikitin

PhD Student
Brandeis University

Ruslan Nikitin is a PhD student at Brandeis University, studying the effects of social networks/ relationships on health outcomes of older adults.