A randomized trial of three diabetes registry implementation strategies in a community internal medicine practice

BACKGROUND: Disease registries are powerful tools with the potential to transform the way chronic diseases are managed. To date, however, little work has been done to determine how to optimize the implementation of a chronic disease registry in practice. METHODS: Twenty-nine physicians and their nurse teams in a large community internal medicine practice participated in this 6-month prospective randomized trial in 2000. Teams were assigned to one of three implementation strategies using information from a diabetes registry. Process and outcome measures for diabetes management were analyzed. Process measures included the percentage of patients completing glycosylated hemoglobin (Hgb) testing within 6 months and low-density lipoprotein (LDL) testing within 12 months. Outcome measures included the percentage of patients with a glycosylated Hgb > 9.3% (equivalent to a HgbA1c > 8.0%), the percentage of patients with an LDL cholesterol > 130 mg/dl, and the percentage of patients with controlled blood pressure, defined as 130/85 millimeters of mercury. Mean change in LDL and glycosylated Hgb values was also measured. RESULTS: Teams randomized to an intervention strategy that included direct letters to patients showed significant improvement across a number of measures. The improvement was most apparent among patients without recent testing or with poorly controlled disease. The two interventions that did not include direct patient letters resulted in limited improvement. DISCUSSION: Disease registries can be used to improve outcomes in the management of diabetes and other chronic diseases. Better outcomes were seen in patients who received letters based on registry-generated data. This strategy should be included as part of a comprehensive chronic disease management plan. Further refinements in the use of registries should result in further incremental improvement.