Evaluation of a Behavior Support Intervention for Patients With Poorly Controlled Diabetes

Background: Disease management programs that include ongoing telephone support for patients with diabetes have shown promise, but published studies have enrolled few socially and economically disadvantaged patients. Methods: We conducted a randomized controlled trial with 201 patients with poorly controlled type 2 diabetes mellitus (72% African American or Latino; 74% with incomes of <=$15 000). Participants were randomized to an intervention package consisting of a 24-minute video behavior support intervention with a workbook and 5 sessions of telephone coaching by a trained diabetes nurse or a 20-page brochure developed by the National Diabetes Education Program. Study measures were completed at baseline, 1 month, and 6 months. Participants' review of the intervention materials was assessed at 1 month. The primary trial end point was hemoglobin A(1c) value. Secondary end points included lipid levels, blood pressure, diabetes knowledge, and self-care behaviors. Data were analyzed with repeated measures analysis of variance. Results: Most participants in both groups (94%) reviewed the intervention provided, and 73% of participants assigned to the experimental group completed 5 sessions of telephone coaching. There was a significant overall reduction in mean (SD) hemoglobin A(1c) value from baseline (9.6% [2.0%]) to 6 months (9.1% [1.9%]) (P < .001), but differences between groups were nonsignificant. Differences on other clinical measures (lipid levels and blood pressure) and measures of diabetes knowledge and self-care behaviors were also nonsignificant. Conclusions: There was no significant effect of the experimental intervention compared with the control condition. The dose of intervention provided was less than in previously published studies. More intensive interventions may be necessary for the most disadvantaged patients.