Date of Award


Document Type

Open Access Thesis


Epidemiology and Biostatistics



First Advisor

Kellee White



The objective of this study is to investigate the association between illness-related diabetes social support (IRDSS) and glycemic control (GC) among a racially diverse sample of middle aged and older adults. In gender-stratified models we examine whether the relationship between IRDSS and GC is modified by race/ethnicity. Additionally, we explore the association between eight individual types of IRDSS and GC.


We used the Health and Retirement Study (HRS) 2002 and the 2003 Diabetes Survey. The final analytic sample consisted of 914 study respondents. Gender-stratified multivariable logistic regression models were used to calculate odd ratios (ORs) and 95% confidence intervals (CIs) for the association between IRDSS and GC, test an interaction between race/ethnicity and IRDSS, and investigate the association between individual IRDSS variables and GC. Final models were adjusted for socio-demographic characteristics, lifestyle factors and clinical characteristics.


79.5% and 76.7% of males and females respectively are classified as having adequate glycemic control. Among females, there was a significant association between IRDSS and GC (OR: 2.39; 95% CI: 1.44, 4.00), but no significant association was observed among males. There was a significant interaction between race/ethnicity and IRDSS (p-value < 0.10). Among Hispanic females, there was a higher odds (OR: 5.75; 95% CI: 1.01, 32.63) of adequate glycemic control in those who had high levels of IRDSS. The strongest association was found between high levels of emotional oriented social support and GC among females (OR=2.05; 95% CI: 1.18, 3.54).


Our findings suggest that the relationship between IRDSS and GC varied by gender, where we only observed an association among females. Although we found a significant interaction between Hispanic race/ethnicity and IRDSS among females, our results should be interpreted with caution due to small sample sizes for the groups. We found that type of social support for self-care activities may play a role in achieving GC, where emotional support was most important. Further studies should aim at identifying factors which may determine gender differences in the association between IRDSS and GC. Race/ethnicity as an effect modifier of this association also needs further exploration in studies with larger sample sizes.

Included in

Epidemiology Commons