Background: Diabetes complications account for significant worldwide morbidity and mortality. Improving glycemic control decreases microvascular complications, particularly among patients with the worst control. Current performance measures fail to prioritize such individuals. The categorization of glycemic control within a safety-net clinic population was compared using a common performance measure against one derived from a metric accounting for change in glycated hemoglobin (A1c) over time.
Methods: Retrospective cohort analysis of all patients in a safety-net primary care clinic population quality registry with confirmed diabetes mellitus who had at least two A1c values between 2007 through 2011. Patients were stratified into five groups' on the basis of maximum and earliest A1c level (< 7%, 7% to < 8%, 8% to < 9%, 9% to < 10%, and > 10%). The change in Alc was assessed over time and compared with standard healthcare effectiveness data and information set (HEDIS) performance measures.
Results: Some 1122 patients were included in the analysis, with mean A1c of 7.9%. There was a modest annual decrease in the average A1c, and > 19% of patients improved by 1% or more during each of the previous three years. For patients who had maximum A1c values > or = 10%, there was a significantly greater reduction in A1c (p < .01), which was not reflected in the standard performance measure.
Conclusions: It is feasible for safety-net clinics to analyze their patients with diabetes by level of disease control on the basis of change in A1c over time. Patients with the worst glycemic control tend to have the greatest improvement but are often overlooked by conventional performance measures. Improved performance measures should focus on longitudinal diabetes control and emphasize reducing risk of complications among patients at highest risk.