878-P: The Association between Demographics, Insulin Rationing, and Glycemic Control Assessed by Hemoglobin A1c

Abstract

Introduction: Insulin cost continues to rise, making affordability difficult for patients. This study looks at the relationship between demographic variables, insulin rationing and glycemic control. Methods: Patients with a diagnosis of type 2 diabetes who use insulin in a resident primary care practice were approached to participate. 131 surveys were conducted by phone (two patients declined), with questions about demographics, medical history, and insulin rationing. Data was summarized using frequency and percentage. Chi-squared/Fisher’s exact test was performed to evaluate the relationship between insulin rationing and variables including age, gender, race, insurance, smoking status, duration of diabetes and insulin use, insulin type, and history of hypertension, hyperlipidemia, obesity, cardiovascular disease, anemia, malignancy, or autoimmune disease. Logistic regression analysis was performed to determine if there is relationship between insulin rationing and hemoglobin A1c (HbA1c) greater than or equal to 8% (Yes/No). Results: Race, gender, smoking status and diabetes duration were associated with insulin rationing. Specifically, black and Asian (vs. white) (p=0.012), males (p=0.049), smoking (yes) (p=0.032) and longer diabetes duration (more than 10 years) (p=0.025) were more likely to ration insulin. Insurance type (p=0.054), age (p = 0.419), and duration of insulin use (p=0.804) were not associated with rationing. Interestingly, comorbidities that confound diabetes control did not affect rationing. There was no association between insulin rationing and HbA1c values greater than or equal to 8% (p = 0.205, Odds ratio 0.37, 95%CI = 0.08 to 1.72). Conclusions: This study did not show an association between insulin rationing and glycemic control, but demonstrated an association with rationing in demographic variables like race and gender. The results of this study were limited by sample size, cross-sectional design and possibly response bias. Disclosure T. Menon: None. M. Zhang: None. A. Myers: None.