Variations in diabetes care and the influence of office systems

Although a variety of office tools have been shown to improve diabetes care, the use and effectiveness of these tools outside the context of clinical trials is largely unknown. We surveyed primary care practices in Kansas. Using Medicare claims, we identified patients with diabetes, linked them to practices, and assessed the use of glycohemoglobin (GHb) tests, eye examinations, and lipid profiles. Performance rates among practices ranged from 15% to 100% (mean = 84%) for GHb tests, 20% to 100% (mean = 70%) for lipid profiles, and 50% to 100% (mean = 78%) for eye examinations. None of the practices used computerized tracking systems, 29% used diabetes flowsheets, and 33% allocated specific diabetes care responsibilities to nonphysician personnel. Neither use of flowsheets nor use of nonphysician personnel was associated with improvements in diabetes care. Although practices varied widely in their care of patients with diabetes, as currently implemented, office system tools have not led to consistently better performance.