Online Program

Trends Over Time in Glucose Control for Type 2 Diabetes Patients with Good or Poor Initial Hba1c Levels: A Multilevel Growth Model

*Upali W Jayasinghe, Centre for Primary Health Care and Equity 
Mark F Harris, Centre for Primary Health Care and Equity 

Keywords: HbA1c; multilevel analysis; type 2 diabetes; longitudinal data; general practice; regression to the mean

Background: To examine within and between patient changes in HbA1c and to investigate whether patients have different trajectories based on the initial HbA1c value. Methods: Multilevel growth modeling was used to explore the changes in HbA1c by allowing each individual in a cohort of 581 patients treated by up to 50 general practitioners in a rural Division of General Practice in Australia to have a unique trajectory over the four years after adjustment for time-varying risk factors and individual demographic variables using annual clinical evaluation data from Division based electronic diabetes register system. Results: Higher levels of HbA1c were associated with higher triglyceride levels, longer duration of diabetes, and younger age. The multilevel model showed that on average, patients who started at a lower level of HbA1c tended to increase HbA1c levels and patients who started at a higher level tended to decrease their HbA1c levels over the years. These effects were not due to regression toward the mean. Conclusions: This study shows the importance of monitoring of HbA1c levels on regular basis and preventing patients with initially good HbA1c levels from deteriorating over time. The results suggest that patients at different levels of HbA1c may benefit from differing interventions. The results of this study are particularly useful for clinicians performing follow-up of type 2 diabetes patients.