1397-P: Characterization of Adults with Newly Diagnosed Type 2 Diabetes by Levels of Early Glycemic Control

Abstract

Glycemic control for type 2 diabetes (T2D) in the US remains suboptimal, and the trajectory of improvement has plateaued over the past decade. The ability to use electronic health record (EHR) data available at the time of diagnosis to proactively risk-stratify newly diagnosed individuals could enable the targeting of more tailored, intensive initial care to high-risk individuals starting on day 1. As a first step we sought to explore the differences between patients by early sub-optimal glycemic control. Using data from a large healthcare system, we identified adults (21-74 years) diagnosed with T2D from 2010-2016. We use EHR-derived patient factors including clinical characteristics, sociodemographics, care utilization patterns, and health behaviors (assessed at T2D diagnosis). We defined suboptimal glycemic control using the time weighted HbA1c over the 5-years post diagnosis. We dichotomized people into those with a time weighted average HbA1c ≥ 8% vs. <8%. We used chi-square tests and t-tests to compare patient factors by level of glycemic control. Among 49,213 individuals, 6,677 (13.7%) had suboptimal glycemic control. Those with suboptimal glycemic control were more likely to be younger (mean age 47.5 vs. 55.1, p<0.001), male (60.3% vs. 50.9%, p<0.001), and Latino (32.1% vs. 20.8%, p<0.001), but less likely to have any prior HbA1c measurements (7.5% vs. 23.9%, p<0.001), an influenza vaccine during the year prior to diagnosis (39.1% vs. 58.3%, p<0.001), and to have a recent primary care provider (PCP) visit (88.3% vs. 95.2%, p<0.001). There were clear demographic and care utilization differences between optimal vs. suboptimal glycemic control which suggests patient and provider engagement may be important for achieving glycemic control. In future work we seek to determine if models leveraging a set of EHR-derived patient variables can predict suboptimal glycemic control. Such risk stratification can support healthcare systems in tailoring the intensity of initial T2D care. Disclosure C.Board: None. S.Alexeeff: Research Support; NIH - National Institutes of Health. A.J.Karter: None. R.W.Grant: None. A.Gopalan: Research Support; American Diabetes Association, National Institute of Diabetes and Digestive and Kidney Diseases. Funding National Institutes of Health (1R21DK130018-01)