Novel algorithms could help individualize HbA1c goals

Two novel algorithms to individualize glycemic control targets would allow clinicians to set HbA1c goals, reclassify from one-quarter to one-third of patients as controlled instead of uncontrolled, and possibly improve treatment of certain patient populations, a study found.

Researchers conducted a cross-sectional, observational study of 12,199 adult patients with diabetes in an 18-practice primary care network affiliated with an academic medical center, measuring HbA1c during calendar year 2011. The algorithms applied 4 patient factors—age, duration of diabetes, presence of macrovascular or microvascular complications, and Charlson comorbidity score—to set 1 of 3 HbA1c goals: <6.5%, <7.0%, or <8.0%. Each patient's HbA1c was compared with these targeted goals and to the standard goal of <7%.

Under both algorithms, patients 75 years and over were assigned a goal of <8%. Looking at duration of diabetes, having been diagnosed within 5 years contributed (but was not sufficient) to having an assigned goal of <6.5%, while patients with long duration (15 years under the first algorithm, 10 years under the second) had a goal of <8%. Any macrovascular or advanced microvascular conditions also led to a goal of <8%. Finally, significant comorbidity (Charlson score of <6 under the first algorithm, <4 under the second) put patients in the <8% group. Results were published by Diabetic Medicine on April 11.

The first algorithm assigned 23.8% of patients a goal HbA1c of <6.5%, 29.7% a goal of <7.0%, and 46.5% a goal of <8.0%. The second algorithm assigned 17.7% of patients a goal HbA1c of <6.5%, 14.4% a goal of <7.0%, and 67.9% a goal of <8.0%. Overall, 55.7% of patients were considered controlled under the standard approach, 61.2% were considered controlled using the first algorithm, and 67.5% were considered controlled under the second algorithm.

The authors noted that the algorithms combine multiple sources of readily available patient information to provide a starting point for decision-making about glycemic goals, although clinicians in practice should consider factors that were not available in this electronic health record-based study, including risk and riskiness of hypoglycemia, social circumstances, and patient preferences. They also acknowledged that the 6.5% goal is controversial and should be based on the interventions required to achieve it or could be left out of the system entirely.

However, the lowered goal could improve existing disparities in diabetes outcomes affecting minority patients. “As a result of younger age and less co-morbidity, individualized algorithms would reclassify slightly more non-Hispanic black and Hispanic patients to more stringent goals. Because these groups have increased prevalence of diabetes complications and diabetes mortality, partially attributable to inadequate glycemic control, the targeted algorithms may help focus attention on glycemic control early in the disease course,” the authors noted.

Overall, the algorithm-based approach could help health systems meet performance requirements without overtreating by reclassifying a quarter to a third of patients from uncontrolled to controlled without adversely affecting vulnerable subgroups, the authors concluded.